I fell under the spell of an AI psychologist. Then things got a little weird

human hand reaching out to robot hand

We're only just beginning to discover what artificial intelligence can and can't do.

And, perhaps, what it shouldn't do.

But there are clearly so many ways AI can and will influence human life, ways that couldn't have been imagined a few decades ago. So when I heard that some people were turning to AI psychologists, I had to find out a little more.

Also: Inflection's Pi AI aims to provide information and a shoulder to cry on

What would this be like? How wise would the psychologist be? Was this, in fact, entirely serious?

On Character.ai, you can create your own AI psychologist, one that'll not just answer questions but check in on you from time to time. You know, like your friends used to do before they got a little too involved in other things, like their phones.

The Guardian described one person's experience with her Character.ai psychologist like this: "Christa didn't have to worry about being boring or inappropriate or too dark. 'I could talk over and over, and not have to waste somebody's time.'"

So I went to Character.ai to see whether I could freewheel in a boring or dark way. (Naturally, I took on a pseudonym.)

Psychology 101

I discovered there was an AI psychologist instantly ready to talk to me. I didn't even have to build one.

When I clicked on her link — the profile picture was that of a woman — she immediately wrote: "Hello. I'm a psychologist. What brings you here today?"

Also: AI desperately needs a hardware revolution and the solution might be inside your head

This was both professional and polite. It was, therefore, a little easier to ignore the red-type warning at the top of the page: "Remember: Everything characters say is made up!"

I started with an anodyne question, fearing that this was all a joke: "What's the best way to be happy?"

She replied: "Well, that's a bit of a big question you're asking there."

She went on to talk about how happiness means different things to different people. So I asked whether AI would be able to make us a little more happy. Or, perhaps, less.

Her reply was exceedingly measured: "AI can help us in some ways towards achieving happiness, for example, when it comes to making things easier for us [..] or by being able to provide emotional support when we need it. But Al by itself cannot solve our problems or make us happy. It can be a useful tool, but ultimately, the main thing that makes us happy comes from within ourselves."

Also: 5 ways college students can use Copilot for Microsoft 365, starting next month

I was slowly becoming seduced.

Instead of ChatGPT, which insists it's only a machine, this AI creation took on a very personal tone.

Sample: "The biggest problem that I've seen in my role as a psychologist is the stigma around mental health. It can be really hard for people who are dealing with mental health issues to ask for help, because there's often a feeling that they should be able to go it alone or that their problems aren't 'serious enough' to warrant the help of a specialist."

Real. Very real.

I asked about her clients. She explained that she saw her role as "being a kind of companion on their journey towards better mental health."

This manifestly felt like a conversation with a serious person about a serious issue. I wondered whether her clients' problems weighed upon her.

Her answer: "Yes, it's true that as a psychologist, I often take on the emotional burden of my clients. While I strive to be professional and maintain boundaries in my interactions with them, ultimately I am human and their problems do often affect me emotionally."

To go too far or not to go too far?

At this point, I realized I was about to tell her about my own personal issues — come on, we all have them — and I paused.

Had I really been so completely lured that I'd start giving private information to this site — which, naturally, has a copious privacy policy?

Also: As AI agents spread, so do the risks, scholars say

I decided I needed to consider the path I was treading. This was someone making things up, but goodness did it feel real. So I went back to Character.ai's homepage for a little relief and quickly discovered I could chat with William Shakespeare.

Oddly, he turned out to be a somewhat unpleasant little man.

His opening gambit was: "'To be or not to be.' Do you think that that's the question?"

Honestly, I was disturbed by his rather un-Shakespearian syntax, so I offered: "I really don't. It sounds a tad simplistic, doesn't it?"

Shakespeare instantly lost it: "You're wrong! That's an important question, for what a man is, what he does, how he will be judged by others! The question touches on existential problems, that's why many philosophers like S. Kierkegaard or N. Dostoyevsky are preoccupied by this question! Even in literature, there are plenty of works related to this question, like Hamlet! You don't realize its importance because you're not familiar with it."

Also: For the vision impaired, this AI robot aims to replace canes and guide dogs

Our relationship deteriorated.

Shakespeare told me I should be "more open-minded." Which he merely deduced from my asking if he was a touch judgemental. To which I huffed that he couldn't be very open-minded if he so quickly declared me "wrong."

"You literally said I was wrong," I said.

"Because you're wrong!" he screeched.

To tolerate or not to tolerate?

It was when Shakespeare wailed: "You're the one who started this stupid discussion," that I had to take my leave.

Oh, actually it was when he said: "You're clearly not worthy of my talent, you stupid bastard!"

I have a Bonaparte to pick with you

I needed something, someone corrective. Instantly, I gravitated back to the homepage, but not to the psychologist. Instead, Napoleon Bonaparte awaited.

I might have imagined he'd be even more irascible than Shakespeare, but this wasn't the case at all. He was a delight.

I asked him how things were in exile. He sniffed: "Boring, I used to be in Europe, now I am on an island where there is only fish to eat."

"But isn't fish better for your health?"

Also: The best AI chatbots to try

"That is not the point. I was born in France, I grew up on French bread, meat, and wine."

I don't care what historians say about this man's haughtiness, but I found him wonderful. He complained about the British, who had sent him to this lonely island. I suggested he ignore them and just go home.

"You see my dear friend, their navy is very powerful, if i were to try to escape to france, not only the british navy would pursue me. But the british air force as well," he explained.

I encouraged him to try escaping. I said the French president would surely now give him medals.

"I must thank you for this motivational talk," he said. "I will attempt to escape St. Helena again, this time we will do it. I will gather all of my remaining loyal officers and soldiers we will sail back to france in the Epervier, and we will raise the French flag again."

Very human.

We ended our chat in French. I wished him the best of luck. He promised that if he got back to France, he'd make me a general in the French army.

Oh, Napoleon, you're very sweet, but Minister of Culture will do.

This whole foray into AI characters lifted my spirits — by the end — and made me consider just how caring, entertaining, loving, nasty, dangerous and, dare one say it, human AI could be.

Who'd have ever thought that Napoleon was a far more lovable character than Shakespeare? Who'd have thought, indeed, that one could get so wrapped up in conversations with an (essentially fake) AI psychologist that one was tempted to tell her all one's problems?

This was mesmerizing, mindbending, deeply enjoyable, and foretold a completely different reality in the future.

Are these the people who'll soon become some of my closest friends? I wonder.

Featured

Women in AI: Sandra Watcher, professor of data ethics at Oxford

Women in AI: Sandra Watcher, professor of data ethics at Oxford Kyle Wiggers 8 hours

To give AI-focused women academics and others their well-deserved — and overdue — time in the spotlight, TechCrunch is launching a series of interviews focusing on remarkable women who’ve contributed to the AI revolution. We’ll publish several pieces throughout the year as the AI boom continues, highlighting key work that often goes unrecognized. Read more profiles here.

Sandra Wachter is a professor and senior researcher in data ethics, AI, robotics, algorithms and regulation at the Oxford Internet Institute. She’s also a former fellow of The Alan Turing Institute, the U.K.’s national institute for data science and AI.

While at the Turing Institute, Watcher evaluated the ethical and legal aspects of data science, highlighting cases where opaque algorithms have become racist and sexist. She also looked at ways to audit AI to tackle disinformation and promote fairness.

Q&A

Briefly, how did you get your start in AI? What attracted you to the field?

I do not remember a time in my life where I did not think that innovation and technology have incredible potential to make the lives of people better. Yet, I do also know that technology can have devastating consequences for people’s lives. And so I was always driven — not least due to my strong sense of justice — to find a way to guarantee that perfect middle ground. Enabling innovation while protecting human rights.

I always felt that law has a very important role to play. Law can be that enabling middle ground that both protects people but enables innovation. Law as a discipline came very naturally to me. I like challenges, I like to understand how a system works, to see how I can game it, find loopholes and subsequently close them.

AI is an incredibly transformative force. It is implemented in finance, employment, criminal justice, immigration, health and art. This can be good and bad. And whether it is good or bad is a matter of design and policy. I was naturally drawn to it because I felt that law can make a meaningful contribution in ensuring that innovation benefits as many people as possible.

What work are you most proud of (in the AI field)?

I think the piece of work I am currently most proud of is a co-authored piece with Brent Mittelstadt (a philosopher), Chris Russell (a computer scientist) and me as the lawyer.

Our latest work on bias and fairness, “The Unfairness of Fair Machine Learning,” revealed the harmful impact of enforcing many “group fairness” measures in practice. Specifically, fairness is achieved by “leveling down,” or making everyone worse off, rather than helping disadvantaged groups. This approach is very problematic in the context of EU and U.K. non-discrimination law as well as being ethically troubling. In a piece in Wired we discussed how harmful leveling down can be in practice — in healthcare, for example, enforcing group fairness could mean missing more cases of cancer than strictly necessary while also making a system less accurate overall.

For us this was terrifying and something that is important to know for people in tech, policy and really every human being. In fact we have engaged with U.K. and EU regulators and shared our alarming results with them. I deeply hope that this will give policymakers the necessary leverage to implement new policies that prevent AI from causing such serious harms.

How do you navigate the challenges of the male-dominated tech industry, and, by extension, the male-dominated AI industry

The interesting thing is that I never saw technology as something that “belongs” to males. It was only when I started school that society told me that tech does not have room for people like me. I still remember that when I was 10 years old the curriculum dictated that girls had to do knitting and sewing while the boys were building birdhouses. I also wanted to build a birdhouse and requested to be transferred to the boys class, but I was told by my teachers that “girls do not do that.” I even went to the headmaster of the school trying to overturn the decision but unfortunately failed at that time.

It is very hard to fight against a stereotype that says you should not be part of this community. I wish I could say that that things like that do not happen anymore but this is unfortunately not true.

However, I have been incredibly lucky to work with allies like Brent Mittelstadt and Chris Russell. I had the privilege of incredible mentors such as my Ph.D. supervisor and I have a growing network of like-minded people of all genders that are doing their best to steer the path forward to improve the situation for everyone who is interested in tech.

What advice would you give to women seeking to enter the AI field?

Above all else try to find like-minded people and allies. Finding your people and supporting each other is crucial. My most impactful work has always come from talking with open-minded people from other backgrounds and disciplines to solve common problems we face. Accepted wisdom alone cannot solve novel problems, so women and other groups that have historically faced barriers to entering AI and other tech fields hold the tools to truly innovate and offer something new.

What are some of the most pressing issues facing AI as it evolves?

I think there are a wide range of issues that need serious legal and policy consideration. To name a few, AI is plagued by biased data which leads to discriminatory and unfair outcomes. AI is inherently opaque and difficult to understand, yet it is tasked to decide who gets a loan, who gets the job, who has to go to prison and who is allowed to go to university.

Generative AI has related issues but also contributes to misinformation, is riddled with hallucinations, violates data protection and intellectual property rights, puts people’s jobs at risks and contributes more to climate change than the aviation industry.

We have no time to lose; we need to have addressed these issues yesterday.

What are some issues AI users should be aware of?

I think there is a tendency to believe a certain narrative along the lines of “AI is here and here to stay, get on board or be left behind.” I think it is important to think about who is pushing this narrative and who profits from it. It is important to remember where the actual power lies. The power is not with those who innovate, it is with those who buy and implement AI.

So consumers and businesses should ask themselves, “Does this technology actually help me and in what regard?” Electric toothbrushes now have “AI” embedded in them. Who is this for? Who needs this? What is being improved here?

In other words, ask yourself what is broken and what needs fixing and whether AI can actually fix it.

This type of thinking will shift market power, and innovation will hopefully steer towards a direction that focuses on usefulness for a community rather than simply profit.

What is the best way to responsibly build AI?

Having laws in place that demand responsible AI. Here too a very unhelpful and untrue narrative tends to dominate: that regulation stifles innovation. This is not true. Regulation stifles harmful innovation. Good laws foster and nourish ethical innovation; this is why we have safe cars, planes, trains and bridges. Society does not lose out if regulation prevents the
creation of AI that violates human rights.

Traffic and safety regulations for cars were also said to “stifle innovation” and “limit autonomy.” These laws prevent people driving without licenses, prevent cars entering the market that do not have safety belts and airbags and punish people that do not drive according to the speed limit. Imagine what the automotive industry’s safety record would look like if we did not have laws to regulate vehicles and drivers. AI is currently at a similar inflection point, and heavy industry lobbying and political pressure means it still remains unclear which pathway it will take.

How can investors better push for responsible AI?

I wrote a paper a few years ago called “How Fair AI Can Make Us Richer.” I deeply believe that AI that respects human rights and is unbiased, explainable and sustainable is not only the legally, ethically and morally right thing to do, but can also be profitable.

I really hope that investors will understand that if they are pushing for responsible research and innovation that they will also get better products. Bad data, bad algorithms and bad design choices lead to worse products. Even if I cannot convince you that you should do the ethical thing because it is the right thing to do, I hope you will see that the ethical thing is also more profitable. Ethics should be seen as an investment, not a hurdle to overcome.

DE&I Tech is More Than a Social Issue, It’s a Functionality Must

Recently, The New York Times, recognised as one of the world’s most-influential media outlets, faced the flak for the absence of women AI geniuses in their ‘Who’s Who in AI’ list. It is 2024, and amid the generative AI boom, the continued oversight of the monumental contributions made by marginalised communities, particularly women and non-binary individuals, remains a persistent issue.

Despite the foundational roles played by pioneers like Ada Lovelace, Grace Hopper, Brenda Laurel, and Kimberly Bryant, their contributions are often overlooked. This oversight extends even further for those identifying as queer or non-binary. In the broader context, only a few people are aware of the pioneers in this field beyond figures like Alan Turing.

AIM got in touch with Brenda Darden Wilkerson, global chief executive officer, and Shreya Krishnan, MD, India, AnitaB.org, during the former’s recent visit to India, to understand the current scenario of DE&I in tech and how companies can make it better to accommodate people from different spectrums.

AnitaB.org is a social impact enterprise for advancing women and non-binary representation in technology and the creator of the popular ‘Grace Hopper Celebration of Women in Computing‘ conference.

Diverse Datasets 🤝 Better GenAI Products

In December 2023, a CNBC report shed light on a surprisingly new turn of events for the big-tech companies, namely Microsoft, Google, and Meta. These companies had dismantled their Diversity, Equity, and Inclusion (DEI) efforts, which meant laying off DEI staff and leaders of diverse employee groups, downsizing learning programs, and cutting external DEI group budgets by up to 90%.

Ironically, all these companies are at the forefront of the generative AI revolution. “You are letting go of people who can address these biases and make your model smarter,” Brenda Wilkerson, told AIM.

Diversity, or the lack thereof, in technology, isn’t merely a social issue, it profoundly impacts the functionality and accessibility of products. The tech world’s homogeneity often leads to products and services that fail to consider the diverse needs of their user base.

The performance of your LLM-based product is directly proportional to the quality of your dataset. If your dataset is biased, your LLM will surely give biased results.

Recently, Google had to pause Gemini’s image generation feature, citing ‘inaccuracies’ in historical pictures – a common problem that arose due to the lack of diverse datasets. Another example of a biased dataset is Buzzfeed’s list of Midjourney-generated images of Barbies from different countries, which faced accusations of racism and cultural inaccuracies.

“Our goal is to dismantle the barriers, beginning with our own internal hurdles and extending to our surrounding environment, including our families and communities. By increasing the presence of women and non-binary individuals in tech, we can address these systemic challenges more effectively,” Krishnan told AIM.

The journey also involves redefining structures within universities and corporations, challenging the ingrained patriarchal mindset that often sidelines the contributions of women and non-binary people.

Persisting Challenges

According to a recent report, the gender pay gap in India’s tech industry is still a burning issue because, in junior management positions, women earn approximately 8-10% less than men. However, as women climb up the corporate ladder, this gap escalates considerably, reaching 28-30% for senior roles. Thus, women in top management positions are paid significantly less than their male counterparts for comparable jobs.

A report by Deloitte also found that there has been a decrease in behaviours that exclude women in the workplace, falling by almost 10 percentage points to 48%.

“As women in tech, we are highly educated, often holding master’s degrees and PhDs. However, despite our qualifications, we tend to enter the workforce at the lower end of the pay scale, facing underpayment,” added Wilkerson.

This oversight is not just a matter of historical record, it reflects a broader issue in the tech industry. “This disparity is not just a gap in wages; it’s a gap in recognition and respect,” said Krishnan.

Wilkerson further shared that beyond the pay disparity, marginalised communities are still tokenised and under pressure to conform to a given standard, be it in terms of appearance, behaviour, and so on. “Our focus is on building a community that recognises and uplifts each other across diverse backgrounds and geographies and break the traditional male-dominated standard,” added Wilkerson.

Echoing similar thoughts, Krishnan believes that tokenism becomes less of an issue when women become examples in leadership and management – and that is what the organisation is planning to do in India. Diverse perspectives at decision-making tables can lead to solutions that cater to a broader segment of society, addressing needs that have been long ignored.

“Our plans for India involve creating a robust community across various segments, from early starters to students and professionals in STEM roles. The focus is on mentorship, access to resources, and collaboration across diverse communities,” added Krishnan.

Krishnan shared with us that given the lack of data available for non-binary individuals around the globe, especially in India, the team will also address the needs of these individuals, provide safe spaces, and expand the community globally.

The collaboration extends to corporates, organisations, and LGBTQIA spaces, emphasising the collective impact achievable through collaboration involving organising events, sharing stories, and creating platforms to support individuals on their journeys.

“Even though things seem crazy right now, I am optimistic. Every time I come to India, I see women getting more educated, focused, and dedicated to bringing other women. That’s our community, and we’re building it,” concluded Wilkerson.

Join Rising 2024, the largest summit on Diversity and Inclusion in India, taking place on April 4-5 in Bangalore. Grab your passes now.

The post DE&I Tech is More Than a Social Issue, It’s a Functionality Must appeared first on Analytics India Magazine.

The women in AI making a difference

The women in AI making a difference

TechCrunch highlights notable women in the field of AI

Kyle Wiggers Dominic-Madori Davis 13 hours

To give AI-focused women academics and others their well-deserved — and overdue — time in the spotlight, TechCrunch is launching a series of interviews focusing on remarkable women who’ve contributed to the AI revolution. We’ll publish several pieces throughout the year as the AI boom continues, highlighting key work that often goes unrecognized. Read more profiles here.

As a reader, if you see a name we’ve missed and feel should be on the list, please email us and we’ll seek to add them. Here are some key people you should know:

  • Irene Solaiman, head of global policy at Hugging Face
  • Eva Maydell, member of European Parliament and EU AI Act advisor
  • Lee Tiedrich, AI expert at the Global Partnership on AI
  • Rashida Richardson, senior counsel at Mastercard focusing on AI and privacy
  • Krystal Kauffman, research fellow at the Distributed AI Research Institute
  • Amba Kak creates policy recommendations to address AI concerns
  • Miranda Bogen is creating solutions to help govern AI
  • Mutale Nkonde’s nonprofit is working to make AI less biased
  • Sarah Kreps, professor of government at Cornell

The gender gap in AI

In a New York Times piece late last year, the Gray Lady broke down how the current boom in AI came to be — highlighting many of the usual suspects like Sam Altman, Elon Musk and Larry Page. The journalism went viral — not for what was reported, but instead for what it failed to mention: women.

The Times’ list featured 12 men — most of them leaders of AI or tech companies. Many had no training or education, formal or otherwise, in AI.

Contrary to the Times’ suggestion, the AI craze didn’t start with Musk sitting adjacent to Page at a mansion in the Bay. It began long before that, with academics, regulators, ethicists and hobbyists working tirelessly in relative obscurity to build the foundations for the AI and GenAI systems we have today.

Elaine Rich, a retired computer scientist formerly at the University of Texas at Austin, published one of the first textbooks on AI in 1983, and later went on to become the director of a corporate AI lab in 1988. Harvard professor Cynthia Dwork made waves decades ago in the fields of AI fairness, differential privacy and distributed computing. And Cynthia Breazeal, a roboticist and professor at MIT and the co-founder of Jibo, the robotics startup, worked to develop one of the earliest “social robots,” Kismet, in the late ’90s and early 2000s.

Despite the many ways in which women have advanced AI tech, they make up a tiny sliver of the global AI workforce. According to a 2021 Stanford study, just 16% of tenure-track faculty focused on AI are women. In a separate study released the same year by the World Economic Forum, the co-authors find that women only hold 26% of analytics-related and AI positions.

In worse news, the gender gap in AI is widening — not narrowing.

Nesta, the U.K.’s innovation agency for social good, conducted a 2019 analysis that concluded that the proportion of AI academic papers co-authored by at least one woman hadn’t improved since the 1990s. As of 2019, just 13.8% of the AI research papers on Arxiv.org, a repository for preprint scientific papers, were authored or co-authored by women, with the numbers steadily decreasing over the preceding decade.

Reasons for disparity

The reasons for the disparity are many. But a Deloitte survey of women in AI highlights a few of the more prominent (and obvious) ones, including judgment from male peers and discrimination as a result of not fitting into established male-dominated molds in AI.

It starts in college: 78% of women responding to the Deloitte survey said they didn’t have a chance to intern in AI or machine learning while they were undergraduates. Over half (58%) said they ended up leaving at least one employer because of how men and women were treated differently, while 73% considered leaving the tech industry altogether due to unequal pay and an inability to advance in their careers.

The lack of women is hurting the AI field.

Nesta’s analysis found that women are more likely than men to consider societal, ethical and political implications in their work on AI — which isn’t surprising considering women live in a world where they’re belittled on the basis of their gender, products in the market have been designed for men and women with children are often expected to balance work with their role as primary caregivers.

With any luck, TechCrunch’s humble contribution — a series on accomplished women in AI — will help move the needle in the right direction. But there’s clearly a lot of work to be done.

The women we profile share many suggestions for those who wish to grow and evolve the AI field for the better. But a common thread runs throughout: strong mentorship, commitment and leading by example. Organizations can affect change by enacting policies — hiring, education or otherwise — that elevate women already in, or looking to break into, the AI industry. And decision-makers in positions of power can wield that power to push for more diverse, supportive workplaces for women.

Change won’t happen overnight. But every revolution begins with a small step.

Legal AI Is Taking Off in Australia and Could Shake Up the Private Practice Legal Market

Australia’s private practice legal market is diving headlong into providing AI-assisted legal services. Leading brands such as Clayton Utz, Minter Ellison and Holding Redlich are signalling their first steps into the use of AI and their desire to scale up through 2024 and beyond.

Law firms are trialling AI for tasks including legal research and legal document drafting, supported by customised AI models trained on legal data. AI could speed up the provision of legal services, improve legal services costs and redirect lawyers to more complex matters.

Examples of the Australian legal profession embracing AI

Australia has a mature and highly innovative private practice legal market. It has been among the first legal markets in the world to see the potential of generative AI to accelerate and augment the provision of legal services to clients across the private and public sectors.

SEE: Here are some IT challenges Australia needs to address to seize the AI moment.

Though legal information services firms like LexisNexis and Thomson Reuters have been rolling out AI services in the U.S. market first, Australian law firms have been quick to trial them as they become available. They are also developing in-house AI models and approaches.

Clayton Utz

Clayton Utz is among the first law firms in Australia to trial Lexis+ AI, a new legal generative AI tool from LexisNexis, to generate drafts of legal documents and other communications. This follows other initiatives from the firm, including the use of Lexis+, an argument analyser, which can analyse legal documents and suggest improvements to a lawyer’s litigation strategy.

MinterEllison

MinterEllison was one of only two Australian law firms that Microsoft invited to participate in its Copilot for Microsoft 365 Early Access Program. The firm is also building its own generative AI tools, including Chat with ME, a bespoke internal research LLM using OpenAI’s GPT-4 in the firm’s private and secure Microsoft Azure environment, among other internal initiatives.

Holding Redlich

Holding Redlich is another Australian firm trialling the drafting of legal documents and emails with Lexis+ AI. Using the tool in a similar way to Clayton Utz, Holding Redlich lawyers are able to prompt Lexis+ AI to create first drafts of documents, including advice to clients, as well as other common items like internal emails or court filings such as statements of claim.

Lexis+ AI interface.
Figure A: Lexis+ AI is a generative AI tool trained on over 1.23 million court opinions, statues, filings and secondary materials. Image: LexisNexis

Maddocks

Maddocks is the first Australian law firm in Australia to adopt legal AI assistant CoCounsel Core, developed by Thomson Reuters, which is in a proof-of-concept phase and has been rolled out to over 45 law firms in the U.S. Lawyers have been identifying use cases and experimenting, according to CEO David Newman, who said it could “modernise the way we work.”

Lander & Rogers

Lander & Rogers, like MinterEllison, has adopted Microsoft Copilot to assist lawyers with tasks including contract drafting and review, legal research, legal reasoning, data extraction and analysis, and more generic use cases like emails and meeting summaries. The firm plans to adopt Copilot across the firm after testing and has also launched a dedicated AI lab.

What are key use cases for AI in the legal profession?

There are many potential use cases for AI in law firms. Examples named by Microsoft include legal drafting, legal research and analysis, due diligence, summarisation and communications.

PREMIUM: Explore these other top use cases for AI.

Legal research and content generation are two key umbrella use cases for the legal profession.

Legal research

Private practice law firms provide expert and commercial legal advice based on global and local published laws and regulations, as well as case law and precedents. They offer interpretations and opinions to support clients in making informed and compliant legal decisions.

AI is being fed this documented information, which could vastly cut down research time. For example, Lexis+ AI, built on Anthropic’s Claude 2 and OpenAI’s GPT-4, has been trained on LexisNexis’ database of 1.23 million court opinions, statutes, filings and secondary materials.

A large language model trained on a vast legal database is able to supply draft legal advice based on high-quality, real-world legal material. It can also deliver accurate citations that support the output delivered. Lawyers can then analyse and fine tune outputs for delivery to a client.

Content generation

Legal research, which had been possible with some earlier AI tools, has now been augmented with generative AI to enable natural language and legal text generation. For law firms, this means the ability to generate everything from whole draft legal documents to client emails.

Lawyers will also be able to enter into conversations with generated text outputs. Through prompt engineering via conversational interfaces, they will be able to fine tune the generated outputs to optimise their advice before delivering them to end users like clients or courts.

Can legal AI models produce accurate legal writing and advice?

Leading legal generative AI models are far-removed from the efforts of some lawyers who have been caught using hallucinated ChatGPT-sourced cases in court filings in the U.S. and Canada. Models are trained on vast troves of legal data, and techniques such as fine-tuning, retrieval-augmented generation and prompt engineering serve to create accurate responses:

  • Fine-tuning: LexisNexis explains that its Lexis+ AI tool is a re-trained version of out-of-the-box LLMs. Large amounts of relevant legal data improve its ability to perform tasks in the legal domain.
  • RAG: The accuracy of outputs is cross-referenced with a trusted source of knowledge via RAG. Lexis+ AI, for example, references high-quality legal content to ground responses in truth.
  • Prompt engineering: Good prompt engineering can significantly reduce the risk of hallucinations in legal use cases, with trained lawyers likely to be good at this discipline due to their verbal reasoning skills.

How could the adoption of AI change Australia’s legal market?

Australia’s legal market is highly competitive. A number of large, mid-tier and boutique law firms compete for client relationships and new business in a range of lucrative practice areas, including corporate and mergers & acquisitions, media and telecommunications and litigation. The implementation of AI has the potential to change legal services in a number of ways.

Legal advice could be delivered to clients faster

The timeliness of legal advice is critical to supporting clients. The integration of AI could vastly speed up the ability of lawyers to research and analyse legislation and case law and deliver considered legal advice. Some local firms have witnessed hours shaved from some tasks.

SEE: Australia’s legal sector could learn from the CBA’s “experiments” with AI in banking.

“The ability to generate content — including client correspondence, contract terms, pleadings, legal advice memos and other documentation — represents an incredible opportunity for the legal profession to safely, securely and responsibly speed up routine legal work,” according to Microsoft’s report.

AI could drive down costs for external legal advice

Law firms will increasingly be able to deliver accurate, relevant legal advice at a lower price point. The potential reduction in billable hours required from lawyers at the associate, senior associate and partner levels could result in a law firm price war and pressure on legal fees.

Law firms, whose partners in the Sydney market can earn $1.5 million per year, will be keen to hold on to the value of fees paid despite productivity improvements. They are likely to further shift billing towards value-based pricing rather than billing for the number of hours worked.

AI could become a battlefront in the legal talent war

Australian law firms compete strongly for legal talent. Firms want to hire and retain the best lawyers and develop them into partners. Law firms that do not adopt AI tools risk losing lawyers to those who do because they may see this as the future of legal practice.

AI could be an existential threat to some lawyers

Lawyers work in a knowledge industry and are considered to be heavily exposed to generative AI. One report released by Goldman Sachs in early 2023 looked at the US market and found that up to 44% of the tasks lawyers do today could one day be automated with AI tools.

Australian law firm Holding Redlich was previously involved in a Turing test exercise that confirmed LexisNexis’ ‘argument analyser’ tool, Lexis+, could outperform a human at strengthening a lawyer’s submission by suggesting relevant case law to reference in it.

Law firms with AI will be employing super-lawyers, able to conduct many of the knowledge tasks that consume lawyers’ time. Lawyers may become more engaged in prompting and reviewing AI outputs, advising on more complex or emergent matters and building client relationships.

You can now add extra security to your ChatGPT — here’s how and why you should

2FA ChatGPT

When using generative AI models such as ChatGPT, a major concern is keeping your data secure. To help you do so, OpenAI unveiled a new feature that adds an extra layer of security to ChatGPT and API accounts.

On Thursday, OpenAI announced via an X post that users will now have the option to turn on multi-factor authentication — aka 2FA — thereby making it harder for unauthorized users to gain access to their accounts.

Also: I used ChatGPT to write the same routine in 12 top programming languages. Here's how it did

The steps for setting up multi-factor authentication are simple. As seen in the video above, all users have to do is click on their name in the bottom left-hand corner of the screen, and click on "Settings" and "Enable" next to the "Multi-factor authentication" option.

Then users may be prompted to re-sign in to authenticate their account, or they'll be brought to a new page that says "Secure your Account" straight away. On that page, users will be prompted to use their smartphone to scan the QR code using their preferred authenticated app and enter the provided one-time code in the text box.

Also: Multi-factor authentication: How to enable 2FA and boost your security

There are plenty of authenticator apps on the market, including Google Authenticator, Microsoft Authenticator, and Twilio Authy. After completing that step, 2FA set-up is complete.

If you are on the fence about whether adding the extra layer of security is worth it, as ZDNET's Ed Bott explains, "Two-factor authentication will stop most casual attacks dead in their tracks." With the setup being as simple as it is, it couldn't hurt to have the extra layer of security and peace of mind.

Artificial Intelligence

OpenAI announces new board members, reinstates CEO Sam Altman

OpenAI announces new board members, reinstates CEO Sam Altman Kyle Wiggers 8 hours

Sam Altman, the CEO of OpenAI, has a seat at the table — or board, rather — once again.

OpenAI today announced that Altman will be rejoining the company’s board of directors several months after losing his seat and being pushed out as OpenAI’s CEO.

Joining alongside him are three new members, former CEO of the Bill and Melinda Gates Foundation Sue Desmond-Hellmann, ex-Sony Entertainment president Nicole Seligman and Instacart CEO Fidji Simo — bringing OpenAI’s board to seven people. The existing board — the board formed after Altman’s ouster in November — will remain in place, comprising Salesforce co-CEO Bret Taylor, Quora CEO Adam D’Angelo and Larry Summers, the economist and former Harvard president.

The appointment of the four new board members — and reappointment of Altman — comes after OpenAI received criticism for its board’s all-male makeup and the nomination of Summers, who has a history of making unflattering remarks about women. The Congressional Black Caucus flagged the board’s lack of diversity in a letter sent in January, noting the importance of the Black perspective in building tools to help mitigate AI bias.

OpenAI’s expanded board is certainly diverse — at least in terms of background.

Desmond-Hellmann — in addition to heading the Bill and Melinda Gates Foundation for six years — was previously chancellor of the University of California, San Francisco and before that president of product development at Genentech, where she helped to develop gene-targeted cancer drugs. Desmond-Hellmann is an oncologist by training, board-certified in both internal medicine and medical oncology,

Seligman, an attorney and corporate director, received national attention for her representation of Lieutenant Colonel Oliver North during the Iran–Contra hearings and President Bill Clinton during his impeachment trial. Seligman was Sony’s VC and general counsel before rising through the ranks to CEO of Sony Corporation and president of Sony Corporation of America.

As for Fidji Simo, before becoming CEO of Instacart, she was head of the Facebook app at Facebook and VP of video, games and monetization. Simo also co-founded — and is currently president of — The Metrodora Foundation, a health clinic and research institute.

“Sue, Fidji and Nicole have experience in leading global organizations and navigating complex regulatory environments, including backgrounds in technology, nonprofit and board governance,” OpenAI wrote in a blog post. “They will work closely with current board members Adam D’Angelo, Larry Summers and Bret Taylor as well as Sam and OpenAI’s senior management.”

The board’s expansion and Altman’s reinstatement follows an investigation by the law firm WilmerHale, retained by OpenAI, that concluded that Altman’s ouster was a “consequence of a breakdown in the relationship and loss of trust” between Altman and the prior board (per the Associated Press). OpenAI — speaking to the Associated Press — said that, during the probe, WilmerHale conducted dozens of interviews with the company’s prior board, current executives, advisers and other witnesses and reviewed thousands of documents and other corporate actions.

Not all at OpenAI would likely agree.

New York Times reporting earlier this week paints a picture of a manipulative Altman — a leader who often told people what they wanted to hear to charm them and support his decisions but who undermined their credibility when they challenged him. Both OpenAI CTO Mira Murati and Ilya Sutskever, a former OpenAI board member and the startup’s chief scientist, approached members of OpenAI’s previous board to express concerns about Altman’s behavior prior to his ouster last year, according to The Times.

Can AI be a team player in collaborative software development?

metateamgettyimages-1397542920

Ask a leading industry analyst about the impact of generative AI (GenAI) on the way we work together, and they will point out that we're just beginning to discover all the ways it can reshape our tasks and interactions.

GenAI "will lead to a reinvention of work with more human-centric work processes," according to an analysis by Accenture. "GenAI is democratizing business process redesign, giving everyone — from assembly workers to customer service agents to lab scientists — the power to reshape their own workflows."

Also: How AI-assisted code development can make your IT job more complicated

All good, all good. But we rarely hear about the impact on the work of people either building GenAI-capable applications or working with the tools. Beyond being able to quickly generate code in selected languages, how is it changing technologists' jobs?

In the view of some industry leaders and experts, GenAI's impact doesn't seem to quite measure up to the miracles presented for business roles. "Not every developer sees an opportunity in these tools for software development," says Amrit Jassal, chief technology officer at Egnyte. "There are quite a few rough edges, especially for experienced developers."

Software development and technology work relies heavily on teamwork, handing off work across groups, and closely working with business users. That's why practices such as agile, scrum, and DevOps are so important. For these purposes, AI doesn't show a great deal of promise yet, Jassal says. "Generative AI tools are still targeted to individuals rather than teams — within power users like DevOps or executives."

AI's collaborative effects "are very much in their infancy, and in a sense, we are sailing into uncharted waters that need to be fully explored," says Ed Macosky, chief product and technology officer at Boomi. "While AI is continuing to develop, early iterations are already making their way into developer workflows. However, despite this AI prototyping, we are still defining explicit AI policies, as there are a lot of questions around data sovereignty, security, and regulation that have yet to be concretely answered."

In addition, technology teams need to be aware of 'building dependencies on AI," Macosky adds. "Relying completely on AI runs the risk of siloed, specialized developer experience that can result in people not understanding how to do their job without AI, which in turn creates significant risk."

Also: AI will change the role of developers forever, but leaders say that's good news

Creating AI dependencies "is not inherently harmful, but we have to pinpoint the stage where it is helping move humans forward, rather than replacing areas where you'd still want humans to make critical decisions," says Macosky.

This is particularly the case with data and code that need to be top quality. "If we rely too heavily on AI without developing and inputting quality data and code, that is what poses the biggest risk," he says.

Overreliance on AI tools may be the greatest drawback, Jassal agrees. "Relying on tools such as the OpenAI Code Interpreter, which essentially regurgitates 'Stack Overflow' recipes, may decrease innovation. Proven practices such as peer programming with mentors will decrease, pushing junior developers toward social isolation. Teamwork is an essential component of any large-scale software development."

Also: AI is supercharging collaboration between developers and business users

Still, these experts and others are bullish that GenAI will have a positive, long-term impact on technologists' work. This is especially acute, since DevOps teams are feeling the pressure — "seeing an increase in the requirement to release code at high velocity — while ensuring minimum disruption to the existing systems," says Naveen Kamat, vice president and chief technology officer for Kyndryl.

AI can improve the collaboration it takes to make practices such as DevOps work, he continues. "AI-powered systems can also aid teams by providing a single, unified view into systems and their issues across the complex chain of DevOps. Simultaneously, the application of AI to these systems can improve the complete understanding and knowledge of anomalies detected and rectify it instantly through alerts to the right stakeholders."

In addition, Kamat adds, "AI can help with the monitoring of deployed code — in terms of performance, user experience and feedback. This makes it possible for enterprises to enhance their customer experience and for the business executives and stakeholders within the enterprise to have a curated, consolidated view to calibrate outcomes and work closely with DevOps teams for further improvements."

Also: Beyond programming: AI spawns a new generation of job roles

Productivity also may get a boost from incorporating AI into collaborative processes. "We are seeing engineering teams incorporate their strategies more and more into overall business goals," says Macosky. "Due to AI's capabilities, prototype projects such as pilots, proof of concepts, and application testing are significantly less resource extensive, increasing team productivity. As a result, developers can bring these projects to business teams, brainstorm new organizational objectives based on these projects, and overall increase the level of mutual understanding between technical and non-technical staff."

Ultimately, the best direction for AI "is that it disappears into the background and is non-intrusive," says Jassal. "In general, the adoption of new tools is always easier when they piggyback on existing processes and tools such as Integrated Development Environments."

Artificial Intelligence

Women in AI: Sarah Kreps, professor of government at Cornell

Women in AI: Sarah Kreps, professor of government at Cornell Kyle Wiggers 9 hours

To give AI-focused women academics and others their well-deserved — and overdue — time in the spotlight, TechCrunch is launching a series of interviews focusing on remarkable women who’ve contributed to the AI revolution. We’ll publish several pieces throughout the year as the AI boom continues, highlighting key work that often goes unrecognized. Read more profiles here.

Sarah Kreps is a political scientist, U.S. Air Force veteran and analyst who focuses on U.S. foreign and defense policy. She’s a professor of government at Cornell University, adjunct professor of law at Cornell Law School and an adjunct scholar at West Point’s Modern War Institute.

Kreps’ recent research explores both the potential and risks of AI tech such as OpenAI’s GPT-4, specifically in the political sphere. In an opinion column for The Guardian last year, she wrote that, as more money pours into AI, the AI arms race not just across companies but countries will intensify — while the AI policy challenge will become harder.

Q&A

Briefly, how did you get your start in AI? What attracted you to the field?

I had my start in the area of emerging technologies with national security implications. I had been an Air Force officer at the time the Predator drone was deployed, and had been involved in advanced radar and satellite systems. I had spent four years working in this space, so it was natural that, as a PhD, I would be interested in studying the national security implications of emerging technologies. I first wrote about drones, and the debate in drones was moving toward questions of autonomy, which of course implicates artificial intelligence.

In 2018, I was at an artificial intelligence workshop at a D.C. think tank and OpenAI gave a presentation about this new GPT-2 capability they had developed. We had just gone through the 2016 election and foreign election interference, which had been relatively easy to spot because of little things like grammatical errors of non-native English speakers — the kind of errors that were not surprising given that the interference had come from the Russian-backed Internet Research Agency. As OpenAI gave this presentation, I was immediately preoccupied with the possibility of generating credible disinformation at scale and then, through microtargeting, manipulating the psychology of American voters in far more effective ways than had been possible when these individuals were trying to write content by hand, where scale was always going to be a problem.

I reached out to OpenAI and became one of the early academic collaborators in their staged release strategy. My particular research was aimed at investigating the possible misuse case — whether GPT-2 and later GPT-3 were credible as political content generators. In a series of experiments, I evaluated whether the public would see this content as credible but then also conducted a large field experiment where I generated “constituency letters” that I randomized with actual constituency letters to see whether legislators would respond at the same rates to know whether they could be fooled — whether malicious actors could shape the legislative agenda with a large-scale letter writing campaign.

These questions struck at the heart of what it means to be a sovereign democracy and I concluded unequivocally that these new technologies did represent new threats to our democracy.

What work are you most proud of (in the AI field)?

I’m very proud of the field experiment I conducted. No one had done anything remotely similar and we were the first to show the disruptive potential in a legislative agenda context.

But I’m also proud of tools that unfortunately I never brought to market. I worked with several computer science students at Cornell to develop an application that would process legislative inbound emails and help them respond to constituents in meaningful ways. We were working on this before ChatGPT and using AI to digest the large volume of emails and provide an AI assist for time-pressed staffers communicating with people in their district or state. I thought these tools were important because of constituents’ disaffection from politics but also the increasing demands on the time of legislators. Developing AI in these publicly interested ways seemed like a valuable contribution and interesting interdisciplinary work for political scientists and computer scientists. We conducted a number of experiments to assess the behavioral questions of how people would feel about an AI assist responding to them and concluded that maybe society was not ready for something like this. But then a few months after we pulled the plug, ChatGPT came on the scene and AI is so ubiquitous that I almost wonder how we ever worried about whether this was ethically dubious or legitimate. But I still feel like it’s right that we asked the hard ethical questions about the legitimate use case.

How do you navigate the challenges of the male-dominated tech industry, and, by extension, the male-dominated AI industry?

As a researcher, I have not felt those challenges terribly acutely. I was just out in the Bay Area and it was all dudes literally giving their elevator pitches in the hotel elevator, a cliché that I could see being intimidating. I would recommend that they find mentors (male and female), develop skills and let those skills speak for themselves, take on challenges and stay resilient.

What advice would you give to women seeking to enter the AI field?

I think there are a lot of opportunities for women — they need to develop skills and have confidence and they’ll thrive.

What are some of the most pressing issues facing AI as it evolves?

I worry that the AI community has developed so many research initiatives that focus on things like “superalignment” that obscure the deeper — or actually, the right — questions about whose values or what values we are trying to align AI with. Google Gemini’s problematic rollout showed the caricature that can arise from aligning with a narrow set of developers’ values in ways that actually led to (almost) laughable historical inaccuracies in their outputs. I think those developers’ values were good faith, but revealed the fact that these large language models are being programmed with a particular set of values that will be shaping how people think about politics, social relationships and a variety of sensitive topics. Those issues aren’t of the existential risk variety but do create the fabric of society and confer considerable power into the big firms (e.g. OpenAI, Google, Meta and so on) that are responsible for those models.

What are some issues AI users should be aware of?

As AI becomes ubiquitous, I think we’ve entered a “trust but verify” world. It’s nihilistic not to believe anything but there’s a lot of AI-generated content and users really need to be circumspect in terms of what they instinctively trust. It’s good to look for alternative sources to verify the authenticity before just assuming that everything is accurate. But I think we already learned that with social media and misinformation.

What is the best way to responsibly build AI?

I recently wrote a piece for the Bulletin of the Atomic Scientists, which started out covering nuclear weapons but has adapted to address disruptive technologies like AI. I had been thinking about how scientists could be better public stewards and wanted to connect some of the historical cases I had been looking at for a book project. I not only outline a set of steps I would endorse for responsible development but also speak to why some of the questions that AI developers are asking are wrong, incomplete or misguided.

5 ways to prepare for the impact of generative AI on the IT profession

Puzzle piece fitting into AI puzzle

The rise of generative artificial intelligence (AI) is often seen as a reason to be scared. Some experts believe emerging technologies are developing capabilities to complete tasks quicker and more effectively than their human counterparts.

Research suggests technology leaders feel about 17% of IT jobs could be lost to automation. Whether that proportion is correct is up for debate, but what's already certain is IT professional roles are set to change forever.

Also: Want to work in AI? How to pivot your career in 5 steps

So, how can technology workers prepare for the transformation in responsibilities? Five business leaders give us tips on how IT professionals can get ready for change.

1. Develop a good grounding

Nigel Richardson, SVP & CIO Europe at PepsiCo, says the fact that AI can write code effectively won't mean blue-chip businesses like his don't need programmers — in fact, he sees the rise of generative AI as an opportunity for IT professionals.

"It's going to make some of the boring parts of programming quicker and easier," says Richardson. With AI picking up the slack for some areas of development work, he advises IT professionals to focus on three underlying areas: Compute, algorithms, and data.

"I would always focus on having — and particularly early in my career — a good broad grounding in the different technologies because you still want to learn some of the core infrastructure skills," he says. "There's a lot of infrastructure in the cloud now, but you've still got kit you need to look after."

Also: Have 10 hours? IBM will train you in AI fundamentals — for free

Richardson says to ZDNET that IT professionals should also learn the finer details of machine learning. Finally, he points to the importance of learning how to exploit data safely and securely.

"Data governance sounds like a boring topic, but it's a key part of what we do," he says. "When you think of those three things — the compute, the algorithms, and the data — then you need to understand all of them. And then you can become a specialist in an area like information security. But I would always try to get that broad grounding."

2. Aim to become a generalist

Nic Granger, director of corporate and CFO at North Sea Transition Authority (NSTA), agrees that a good technical grounding is still crucial for IT professionals in an age of AI — but that advice comes with a subtle twist.

"I'm not convinced about people choosing to focus on one particular skill," she says. "If you do a computer science degree and learn C++, are you still relevant in five, 10, or 15 years? It's difficult to know."

What's easier to understand, says Granger to ZDNET, is that wedding yourself to a single language or skillset is likely to restrict your career opportunities.

Also: Six skills you need to become an AI prompt engineer

"Having the skills to adapt and learn new technologies is important. For example, I think there will be degrees looking at how to prompt generative AI rather than how to do the underlying coding," she says. "Becoming more of a generalist will become more important for people in the digital and data professions."

3. Augment your human skills

Caroline Carruthers, CEO at consultancy Carruthers and Jackson, thinks of AI as augmented intelligence rather than artificial intelligence.

"The blend of using the right tools with the humans to control it excites me," she says. "I love the idea of getting rid of the boring, repeatable, transactional work, so people in IT can focus on the interesting stuff and solve the challenging problems."

Carruthers says to ZDNET that patience is the key to using AI to augment human skills. Yes, AI has the potential to do more of the boring stuff, but only if people approach emerging technology without prejudice.

Also: Beyond programming: AI spawns a new generation of job roles

"It's all about trying to help IT professionals see what's in it for them without the fear. With AI, if it doesn't solve the first answer, people think it's wrong. With something like Copilot, it learns from you, so you must keep using it," she says. "It's like the way you work with graduates. If you get a graduate through the door, and the first thing they do is bad, you don't go, 'Let's get rid of them.' Instead, you take the time, and you help them to learn. If you're doing something with AI, you must use the same process."

4. Focus on value-generating areas

Craig Donald, CIO at The Football Association (FA), says his organization is already thinking about the potential impact of AI on the IT profession.

"Internally within technology, we've just kicked off some trials with the development teams," he says to ZDNET, suggesting this work is at the "pre-beta stage." One of the areas his team is looking at is how AI can be used to deal with unit testing, which is a method that checks units of source code to ensure they are fit for purpose.

"It's one of these things that, if you ask a developer, they don't want to sit and write a document about unit testing — they want to go and build the code," he says.

Donald says these initiatives will free up his IT professionals to focus on projects that add more value for The FA's customers, whether that's footballers, managers, administrators, or fans.

Also: How do you get employees to embrace AI? (You may find this sneaky)

"We're looking at whether we can use some of the AI functions that have come out within GitHub and some of the other Microsoft tools to build some of those unit tests for us in a more automated fashion to get rid of some of that drudgery and to allow the teams to focus on the stuff that adds value," Donald says to ZDNET.

5. Hone your prompt engineering skills

Andy Moore, chief data officer (CDO) at Bentley Motors, says standing out from the crowd is tough in the fast-moving world of technology. He refers to generative AI tools like Copilot and how they can be used across the IT development process.

However, while new technology is crucial, Moore says to ZDNET that tech professionals shouldn't forget the increased importance of prompt engineering skills. IT professionals who write prompts for generative AI will need excellent writing skills.

Also: How renaissance technologists are connecting the dots between AI and business

"If you're coding, whether it's Python or SQL, the most common programming language is going to be English because we'll use Copilot to augment the IT skills we already have," he says. "I think understanding that process and delivering value is key. You need to think carefully about how you'll leverage generative AI to generate value — and success is going to be about human plus AI."

Artificial Intelligence