Google is backing these 20 startups to help improve the world with AI

Google logo in New York City Pier 57 location

In 2020, Google launched its Google for Startups Founders Funds, providing equity-free cash awards to Black and Latino entrepreneurs. In 2024, all 20 recipients of the funds are connected to the hottest technology of the moment — artificial intelligence (AI).

This year's winners use AI to help solve real-world problems, from preventing wildfires to reducing home energy costs and usage.

Also: How to sign up for Google Labs — and 5 reasons why you should

The founders will receive $150,000 in non-dilutive cash awards and $100,000 in Google Cloud credits. Money aside, the founders will also receive other help to grow their businesses, such as mental health resources and mentorship from Google experts across AI and sales.

Google said in its announcement that the company created the Founders Fund to help "level the playing field" for Black and Latino entrepreneurs who have difficulty accessing early capital, as racial equity is linked to economic opportunity.

"AI can enable startups to build transformative products and solve complex challenges, but founders need access to capital to realize this potential," Maya Kulycky, VP of strategy and operations at Google Research, said in a statement. "Through the Google for Startups Founders Funds, we are proud to invest in promising Black and Latino founders who are leveraging AI technology to help address some of today's most pressing issues."

Also: Google's Gemini AI chatbot now available to younger students in Workspace — how it's different

Here is the full list of winning startups: Akeptus, Beta Financial Services, Bountiful, Cambio AI, EdVisorly, Elis, Hacware, Hire Henry, Hue, Improving Aviation, InOrbit, JustAir Solutions, Maverick, Pagedip, Raincoat, Sensagrate, Sortile, TackleAI, Trustible, and Waterplan.

You can read more about each startup on Google's website, where you can also learn how to get involved with Google for Startups programs, including accelerators, academies, and funds.

Google

Google Rolls Out Gemini in Gmail Side Panel for Google Workspace Users

Google has announced the general availability of Gemini in the Gmail side panel, extending its capabilities beyond Google Docs, Sheets, Slides, and Drive. The new feature, powered by Google’s advanced models like Gemini 1.5 Pro, enhances productivity for Google Workspace users. Gemini in Gmail allows users to:

  • Summarize email threads
  • Receive suggestions for responses
  • Draft emails with assistance
  • Search for specific information within emails or Google Drive files

The integration aims to streamline email management by providing proactive prompts and enabling freeform questions. Users can ask Gemini to retrieve details such as PO numbers, expenditure on events, or upcoming meetings directly from Gmail without leaving the interface.

In addition to the web version, Gemini is now accessible on the Gmail mobile app for Android and iOS. This mobile functionality includes analyzing email threads, presenting summarized views, and upcoming features like Contextual Smart Reply and Gmail Q&A.

End users of Google Workspace stand to benefit from Gemini’s integration across various applications, facilitating efficient workflows and quick access to information.

The future’s connection with Workspace apps like Docs, Sheets, Slides, and Drive enables smooth collaboration and data retrieval. To utilise Gemini in Gmail, admins must ensure smart features and personalisation are enabled for users.

The rollout schedule for the web version spans from June 24 for Rapid Release domains to July 8 for Scheduled Release domains. Mobile rollout is concurrent, with full availability expected within 15 days for both release types.

Gemini in Gmail is accessible to Google Workspace customers with specific add-ons, including Gemini Business and Enterprise, Gemini Education and Education Premium, and Google One AI Premium.

The post Google Rolls Out Gemini in Gmail Side Panel for Google Workspace Users appeared first on Analytics India Magazine.

Bringing Human and AI Agents Together for Enhanced Customer Experience

Sponsored Content

Softweb Solutions AI Agent

Businesses understand that it is important to not only provide good products or services, but also offer great customer experience. Companies often struggle with high volumes of inquiries, inconsistent service quality, and escalating operational costs.

68% of clients that stop doing business with organizations, do so because of poor on-going customer service, not because of dissatisfaction with the product or service.

– The American Management Association Study

What makes a customer happy?

  • Quick turnaround time: Whether it is receiving their products fast or resolving their queries quicker, customer appreciate prompt response.
  • Better quality products: Customers expect that the product or the service they get is well-made and function as intended.
  • Easy query resolution: No business wants a frustrated customer. Keeping them happy by addressing their issues is key to successful and longer customer relationships.
  • Availability of a customer representative: Sometimes, people just want to talk to a real person. Having a helpful representative available shows the customer the company cares.

How can you ensure all these?

The answer is simple: AI agents. By leveraging AI’s ability and scalability alongside human empathy and understanding, businesses can elevate their customer support operations. This ensures seamless, personalized, and efficient customer experience.

What are AI Agents and how do they Transform Customer Support Process?

AI agents are software programs that can learn and act autonomously. They can handle tasks like answering questions, solving problems, and even making decisions, all to achieve specific goals. AI agents can help you address business problems, such as:

Limited Agent Scalability

It is difficult to manage customer queries during peak hours or unexpected surges in inquires. This can lead to increased wait times, making customers frustrated. This in turn, damages brand reputation.

With AI agents in place, you can efficiently handle customer queries by incorporating natural language processing (NLP). AI agents can deflect significant volume from human agents for basic troubleshooting.

Inconsistency and Knowledge Gap

Often, knowledge base (KB) articles are not readily available or updated frequently. This makes maintaining consistent and accurate information across the support team challenging.

AI agents can help in automating KB creation and updates. This ensures that support agents have all the required information handy for faster resolution.

Sentiment Analysis and Proactive Support

It is a tedious task to manually identify customer sentiment within the support tickets. AI agents enable companies to analyze texts for sentiment analysis. By analyzing text for emotional cues, these tools can flag frustrated customers, enabling proactive intervention from human agents and preventing escalation.

Key features of AI Agents

Softweb Solutions AI Agent

Autonomy

AI agents operate independently and perform the required tasks with minimal human intervention. They are also capable of adaptive learning. This helps in improving the performance of AI agents over time.

Reactivity and Proactivity

AI agents can respond to changes in the environment in real time, life self-driving car. They can also anticipate future needs and take proactive measures.

Interactivity

AI agents can efficiently collaborate with human agents to offer better customer service and achieve complex objectives.

Intelligence

AI agents use ML models and AI algorithms for various use cases. For example, it can help businesses with predictive maintenance reminders, forecast a customer’s next purchase, and suggest upselling opportunities.

What are the Key Components of AI agents?

AI agent architecture refers to the underlying structure that enables an AI agent to perceive its environment, make decisions, and take actions. Here are the key components:

Perception Module

This module enables AI agents to collect information about environment through sensors like cameras, microphones, APIs, etc.

Learning Module

This module enables the agents to improve their performance by learning through different machine learning (ML) models like reinforcement learning, or supervised learning.

Knowledge Representation and Reasoning Module

Agents need to be fed data to draw conclusions and make inferences. With techniques like logic rules, semantic networks, or probabilistic models, agents can gain information about the organization.

Action Selection Module

This module allows the agents to process the information and decide what actions to take. This can involve planning algorithms or decision-making frameworks.

Action Module

This module enables agents to translate the action into real-world commands. In case of a physical robot, it can be controlling motors. In case of software, it can be sending messages.

These components work together in a cycle:

Softweb Solutions AI Agent

Some architectures may also include additional components like a communication module for interacting with other agents and a performance monitoring module for tracking the agent's effectiveness.

How can AI and Human Collaborate to Offer Optimal Customer Support

Implementing a successful AI-human collaboration requires a well-defined approach:

Task Identification and Process Automation

You should consider factors like inquiry frequency, resolution time, and required agent expertise. Pinpoint high-volume, rule-based tasks that you would like to automate.

Selecting the right AI tools

  1. NLP-powered chatbots: Easily integrate chatbots that can understand natural language, answer FAQs, and collect customer data.
  2. ML-powered virtual assistant: Automate tasks like appointment scheduling, ticket routing, and more.
  3. Sentiment analytics tools with text classification capabilities: Classify text sentiments (positive, negative, neutral) and flag frustrated customers.
  4. AI-powered knowledge base management: Automate KB creation, updates, and information retrieval.

Integration and Scalability

Integrate AI tools with existing CRM system and other ticketing tools. Build AI agents with flexible scaling options to accommodate for future demands.

Training and Enablement

Invest in training human agents on AI capabilities, escalation protocols, and effective collaboration practices.

Continuous Monitoring and Improvement

Track key metrics like resolution times, agent workload, and customer satisfaction. Analyze data to identify areas for improvement and refine your human-AI collaboration strategy over time.

Measuring Success and Continuous Improvement: The Power of Human-AI Collaboration in Agent Performance

Defining Success Metrics

Establish clear metrics that align with your AI agent’s objectives. Consider these goals:

  • Efficiency: Reduced resolution times and increased inquiries handled automatically.
  • Accuracy: Improved task completion rates and lower error rates in recommendations or predictions.
  • Customer satisfaction: Higher CSAT scores, positive sentiment analysis in interactions.
  • Cost savings: Reduced operational expenses through automation.
  • Quantitative and qualitative metrics: Measure resolution time along with qualitative insights provided by human agents, such as, customer surveys, and human agent feedback.

Data Collection and Collaborative Monitoring

Agent performance data

  • Input/output logs: Human agents can identify potential areas of improvement.
  • Action logs: Monitor actions taken by agents for collaborative analysis of effectiveness.

User interaction data

  • Clickstream data: Human agents can identify UX issues by analyzing user clicks and navigation.
  • Sentiment analysis: Refine the AI agent's language processing and response capabilities through human feedback about sentiment analysis.

Evaluation and Analysis

Regular reviews:

  • Analyze collected data and assess agent performance against defined metrics. Human agents can contribute valuable insights during these reviews.

Identifying areas of improvement:

  • Accuracy issues: Analyze errors with human agents to pinpoint weaknesses in AI agent's knowledge or reasoning capabilities.
  • Inefficiency: Identify bottlenecks in processing or response times.
  • User experience: Use qualitative data and human feedback to understand user pain points and improve the AI agent's interaction design.

Continuous Improvement and Techniques

  • Model retraining: Retrain the AI model with new data or adjusted algorithm based on the identified issues.
  • Rule-based update: Update the rules with human inputs to address shortcomings and adapt to the changing user demands.
  • Interface refinement: Refine the user interface based on user feedback and collaboration with human agents to improve ease of use and clarity.

Real-World uses of AI Agents

Hilton Hotels

This leading global chain of hotels integrated AI agents to improve guest experience. They deployed AI chatbots to handle basic queries and recommend restaurants to their guests. This allowed the human agents to focus on complex issues that required personalized assistance.

The result:

  • Significant reduction in resolution times
  • Boost in guest satisfaction scores

Berkshire Hathaway HomeServices Fox & Roach

This real estate company based in Philadelphia has implemented an AI agent to increase lead generation opportunities generated every month. They deployed an AI chatbot named Taylor. It helps in effectively interacting with hundreds of website visitors in real time.

The result:

  • Increase online lead generation by more than 71%
  • Create successful personalized nurturing campaigns with a 40% open rate

KLM Royal Dutch Airlines

This oldest global airline uses Dialogflow by Google for their AI agent named Blue Bot (BB). They wanted to provide conversational approach to their flyers. BB enables KLM’s passengers to easily book tickets, offers flying advice, shares weather forecast, and more.

The result:

  • Exceptional booking experiences
  • Faster issue resolution

Leverage the Symbiotic Collaboration of Human and AI Customer Support

By embracing this human-AI partnership and implementing the strategies outlined here, you can transform your support ecosystem. Move beyond simply measuring success – use these insights to propel continuous improvement and unlock the true potential of your AI agents. Invest in the successful collaboration of humans and AI agents; it's the key to unlocking the true potential of your customer support operations.

More On This Topic

  • MiniGPT-4: A Lightweight Alternative to GPT-4 for Enhanced…
  • How Do Data Scientists and Data Engineers Work Together?
  • LangChain + Streamlit + Llama: Bringing Conversational AI to Your…
  • AgentGPT: Autonomous AI Agents in your Browser
  • Why You Need To Know About Autonomous AI Agents
  • The Growth Behind LLM-based Autonomous Agents

Some Open Source Software Licences are Only ‘Open-ish,’ Says Thoughtworks

It has been estimated 90% of organisations use some form of open source software, and if they needed to go and code it again themselves, it would cost USD $9 trillion. This makes open source a huge global economic resource.

However, some tools have shifted to commercial models in recent times. After years of growth through developer contribution and widespread uptake among users, they are monetising the end result — often to the chagrin of developer communities and dependent business users.

Global technology consultancy Thoughtworks identified the trend in its most recent Technology Radar. Australian Chief Technology Officer, Scott Shaw, said it is partially driven by a closer focus on financials in recent times, and organisations need to ensure they approach open source “with their eyes open.”

Some open source favourites have shifted to commercial licences

In April 2024, Thoughtworks noted a “churn in the previously serene landscape” of open source. “Several prominent tools have recently garnered bad press, when their maintainers switched — in several cases abruptly — from an open-source licence to a commercial model,” it said.

The trend has been building for some years, according to Shaw. While the tech industry has a common set of principles and a number of well understood open source licences governed by the Open Source Initiative, there has been a growing “divergence” from that paradigm.

Abrupt changes to open source licences

The first example are those companies that have changed the terms of their open source licence mid-stream. After building a developer community and onboarding large numbers of users who have integrated the software into workflows under the permissive standards of open source licences, there has been a move to clamp down on that, often linked to revenue.

SEE: The 8 best open source project management software for 2024

While Thoughtworks wrote that “we have no problem paying for software and are fine with the common model of commercial licences for additional functionality,” it added that “we find it problematic when core functionality of a widely used tool is suddenly put behind a paywall, especially when an ecosystem has developed around the tool.”

‘Semantic diffusion’ in open source

There has also been a blurring in what open source means, with Thoughtworks observing “software that proclaims to be open source, yet fundamental capabilities only appear after consumers pay subscriptions or other charges.” In some cases, an open source project may only distribute code, not builds, increasing the burden for organisations using it on premise.

“One example is some large language models that are being loosely referred to as open source that are not; they are open in some way, but they don’t meet the principles of open source, certainly not the way the OSI defines them,” Shaw said.

Docker, Terraform and Llama 3 diverge from pure open source

Thoughtworks said there have been several examples of shifts to commercial licences or “open-ish” licences emerging. Three examples are developer containerisation software Docker, Hashicorp’s Terraform, and Meta’s newly released LLM Lllama 3.

Docker

Docker is open source software used by developers to automate the deployment of applications inside containers. It became the basis for most application distribution and integral to software delivery, with 55% of developers using it daily. Docker also had a convenient Docker Desktop, allowing developers to run Docker locally on a machine to perform testing.

In 2021, and effective in 2022, Docker changed its licensing. While remaining free for small businesses with fewer than 250 employees and less than USD $10 million in revenue, larger enterprises using it professionally needed to pay for a Pro, Team or Business membership, meaning organisations were no longer in compliance if they did not pay fees to Docker.

Terraform

Terraform from Hashicorp is one of the most popular and effective infrastructure as code tools for safely and predictably provisioning and managing infrastructure in any cloud. However, Hashicorp caused an outcry in the open source community when it made the decision to shift from a Mozilla Public Licence v2.0 to a Business Source Licence, because of its widespread use as an open source software supporting DevOps operations and companies.

SEE: The 5 best open source CRMs for 2024

The company explained its decision, primarily, as being to protect its interests from competitors using Terraform to compete with Hashicorp, who can now utilise commercial licences. This did not placate the whole open source community; some were galvanized to start OpenTofu, a community-driven project that aims to create a fork of Terraform and maintain it as an open-source tool, in line with the company’s previous commitments to open source.

Llama 3

Meta’s Llama 3 is being received as a powerful LLM model, Shaw said. However, in terms of its open source credentials, the model has open weights but does not follow other OSI principles like the ability to examine source code and complete unrestricted redistribution. Meta’s Llama 3 requires the payment of licensing fees based on user numbers for the use of weights.

“If you ask Meta, they call it an openly available model. That is honest, but the term open source gets very loosely applied to these things, and I think it’s important for people to understand openly available or free doesn’t necessarily imply open source. I think this is sometimes missed; people don’t completely understand what degree of openness a particular model might have.”

AI LLMs come in many degrees of openness

Thoughtworks said “semantic diffusion” of the open source badging is something being seen in the fast-growing AI space in particular. “Even though this business model has existed before, it seems to be exploited more with many of the shiny new AI tools — offering amazing capabilities a little too hidden under the fine print,” the firm wrote in its Technology Radar.

Shaw said that for LLMs, there’s a range of openness available in different models. They range from completely proprietary, like OpenAI’s ChatGPT, to models where the source code, training data, model structure and weights are all freely available and open for inspection and contribution. One recent example is Snowflake’s Arctic LLM, released on an Apache 2.0 licence.

Two reasons why companies rethink open source licences

Thoughtworks suggests revenue and IP protection are behind some of the licensing moves.

Focus on financials

The whole tech industry has been more cost conscious in recent years due to economic headwinds, with chief financial officers becoming more influential in decision making. Thoughtworks’ Technology Radar said “a lot of blame has been placed on private equity and venture capital firms for putting more pressure on firms for revenue and profitability, particularly as the tech industry has slowed.” Shaw said it has been a time where people all through the industry have been re-examining their business models, leading to some churn in open source.

The protection of IP

Another factor, noted by Hashicorp in its Terraform licensing decision, is the protection of IP. Thoughtworks writes that “others speculate that the open source vendors are only protecting themselves and their intellectual property from the cloud vendors who would profit from the IP through hosted cloud services.”

Shaw said in some cases bigger organisations, like hyperscalers, had been taking open source tools and creating very profitable services and not paying and licensing fees back to the originator of the tools. Though that is essentially the spirit of open source, the originating vendors want to ensure that they receive some form of financial benefit.

There are risks for enterprises when open source licences change

When the licences of widely used open source software projects shift to a more commercial model, it creates a “big headache” for their enterprise users, Shaw said. To remain compliant with licensing terms, companies have to make sure the software — such as Docker Desktop, in the case of Docker — is removed from individual devices; otherwise, they may be hit with licence fees or risk getting caught out in an audit, even if the software is still there unwittingly.

Shaw said organisations already spend a lot of time, money and effort auditing, making sure the software their employees are using are being used within the terms of their licences. Abrupt shifts in the deal on offer from open source providers can be difficult to manage. “I think it’s something that boards, CEOs and CFOs would want to be conscious of, because they may be highly dependent on open source software that has changed its licensing terms,” Shaw said.

Things IT should watch when using open source software

Thoughtworks has advised businesses and IT stakeholders to exercise “particular diligence around licence issues. Pay attention to caveats and make sure that all files in a repository are covered by the licence at the top level,” the firm detailed in its Technology Radar. Shaw added that enterprises needed to approach open source software with their “eyes open.”

Check the details of open source projects

One factor to look at is whether an open source project is truly grassroots supported, or is dependent on a commercial interest with no other apparent business model, Shaw said. In the latter case, he recommends considering if it is worthwhile paying for the enterprise version of the software, so the terms of the licensing are agreed upon contractually from the start.

Beware of data leakage to SaaS models

Another factor to consider is whether the open source software is actually running on a desktop or is sending some data to the cloud. Shaw said enterprises should know how data is being treated if it is an online service and what sort of safeguards there are against redistribution. In some cases, Shaw said there is a risk of data leakage if organisations are not careful.

New vendors and products are competing after licencing changes

When an open source tool changes licence terms and users are forced to pay, there are always competitors waiting in the wings to step in and provide competition, Shaw said. For example, in the firm’s Technology Radar where it flags tools to watch, alternatives to Docker Desktop include Colima. And while the current economy is causing closer scrutiny of business fundamentals, those accentuated drivers for shifting to commercial licences may be cyclical.

Stability AI lands a lifeline from Sean Parker, Greycroft

Stability AI, the beleaguered generative AI startup behind Stable Diffusion, has raised new cash. But it won’t reveal how much. Greycroft, Coatue Management, Sound Ventures, Lightspeed Venture Partners, O’Shaughnessy Ventures and angel investors Prem Akkaraju, ex-Google CEO Eric Schmidt, Robert Nelsen and Napster founder and ex-Facebook exec Sean Parker have injected fresh capital into Stability, […]

© 2024 TechCrunch. All rights reserved. For personal use only.

Oracle Unveils Clinical Digital Assistant to Reshape Interactions Between Practitioners and Patients

In our fast-paced healthcare environment, medical professionals are constantly challenged to balance the demands of providing high-quality patient care and completing administrative tasks such as managing electronic health records (EHR), which is a digital version of a patient's medical history and other health data.

We know that as AI and machine learning capabilities keep getting better, the future of healthcare is data-driven. However, the technologies need data to reach their full potential. Healthcare providers have to input patient data into the system. However, the inefficiencies of completing such tasks often result in poor patient care or healthcare provider burnout.

According to research, physicians spend 44.9% of their time on the EHR, with 20.7% on EHR input alone. Clinician burnout is at an all-time high, and one of the reasons for that is the burden of administrative work. The overburdened healthcare provider is also a risk to patient safety and is one of the causes of workforce shortage in the healthcare industry.

Oracle, a global leader in cloud technology and enterprise software solutions, has unveiled an innovative solution this week to address some of these challenges. The Oracle Clinic Digital Assistant combines GenAI, clinical intelligence, integrated dictation, and multimodal voice tools to minimize the need for manual input into EHR. This has the potential to have a significant positive impact on improving patient care and reducing healthcare provider burnout. Oracle’s Clinic Digital Assistant is now generally available in ambulatory clinics in the U.S.

(Jonathan Weiss/Shutterstock)

According to Oracle, the new tool can streamline healthcare workflows, and reduce documentation time by 20-40%. This enables healthcare providers to focus on what matters most — their patients.

Integrated with the Oracle Health Electronic Health Record (EHR), the digital assistant offers conversation-based note generation, clinical automation, and can propose follow-ups directly at the point of care.

The physicians would no longer need to navigate complex drop-down menus or scroll through multiple screens to find and enter patient data. With the Oracle Clinical Digital Assistant, they can immediately access the patient’s medical record through voice commands.

During the appointment, the digital assistant captures data using the healthcare provider’s preferred templates, which are built into the platform. Unlike other solutions, Oracle Clinical Digital Assistant takes only a few minutes to complete note generation for the appointment. In addition, it can propose next-step actions, such as drafting referrals and prescription orders and scheduling follow-up appointments. These features not only help save time through better efficiency but can also improve the accuracy of the EHR data.

Oracle tested the new digital assistant with 13 early adopters, including St. John’s Health, Billings Clinic, Hudson Physicians, T.J. Regional Health, and Covenant Health. The results show that the new tool can save physicians an average of four and a half minutes per patient and 20-40% in documentation time.

According to Patricia Doolin, APRN, T.J. Regional Health, the Oracle tool is saving her “up to 10-12 minutes per patient”. This efficiency allows doctors to spend more quality time with their patients, enhancing overall patient care.

“Oracle Clinical Digital Assistant is the most important EHR technology update that I am going to see in my career. Since the 1990’s, EHRs have turned physicians into keyboard junkies. This will change that,” said James Little, MD, primary care physician, St. John’s Health. “Our physicians who have been using this technology have been able to document their patients’ visits in real-time, allowing them to leave at the end of the day with good, quality notes. Time spent after hours documenting is no longer needed.”

According to Dr. Ryan McFarland, family medicine practitioner, Hudson Physicians. Oracle’s Clinical Digital Assistant has been a “game-changer” as it allows him to focus on his patients during the appointment and reduce the time he spends on updating notes post-appointment and after-hours.

While the digital assistant captures all the data automatically, the users have complete oversight and control. They can review, modify, and approve notes and next-step actions on their computer or mobile device. The solution also synchronizes the appointment date with the patient’s medical record, however, it does not copy and paste it over it. This ensures valuable data doesn’t get overwritten.

Related Items

How AI Will Impact the Drug Discovery Pipeline

IBM, Cleveland Clinic, And Hartree Center Collaborate To Advance Healthcare Through Advanced ComputingNew AI Technique Helps Find Alzheimer’s Drug Targets

New AI Technique Helps Find Alzheimer’s Drug Targets

Gmail users can now ask Google’s Gemini AI to help compose and summarize emails

Gemini AI chatbot

If you want AI-powered assistance when managing your emails in Gmail, you can now call on Google's Gemini. In a blog post published Monday, Google announced that Gemini is now rolling out as an integrated option for Gmail on the web and in the mobile app.

Gemini is available to Google Workspace customers with a Google One AI Premium plan, the Gemini Business and Enterprise add-on, or the Gemini Education and Education Premium add-on. Prior to this rollout, the only way to take advantage of AI features in Gmail was by joining the Google Workspace Lab.

Now, you can ask Gemini to help you respond to an email or summarize an existing email thread. The AI tool can also answer questions and find specific details about emails in your inbox and files saved in your Google Drive.

Also: What is Gemini? Everything you should know about Google's new AI model

In Gmail on the web, Gemini will pop up in a left sidebar nestled next to your emails. In Gmail for iOS and Android, the AI tool should appear in a small window when you tap the subject line of a message. The move follows the recent integration of Gemini in Google Docs, Google Sheets, Google Slides, and Google Drive.

The sidebar or window gets you started by offering certain suggestions, such as "Summarize this email," "List action items from this email," and "Suggest a reply to this email." Selecting an option for More Suggestions displays other tasks, including "Draft a reply," and "Help me reply." If you're having trouble composing an email, an icon for "Help me write" is also available in the new email window.

Beyond using the suggested requests, you can ask Gemini questions directly. As a couple of examples cited by Google, you might ask the AI tool questions such as "How much did the company spend on the last marketing event?" or "When is the next team meeting?" In response, Gemini should be able to retrieve the emails that meet your query.

Google has outfitted Gmail with its latest AI model, Gemini 1.5 Pro. This model boasts hefty improvements over its predecessor, including a longer context window, better understanding, and faster performance. Gemini 1.5 Pro is able to handle as many as 1 million tokens, a jump from Gemini 1.0's 32,000 tokens. Such an increase is significant, because the more tokens a model can handle, the more its responses are likely to be better informed.

Google is rolling out the Gemini integration now. For Gmail on the web, users on the rapid release schedule should see it over the next couple of days while those on the scheduled release will have to wait until July 8. For the Gmail mobile app, rapid release and scheduled release users can expect Gemini to appear over the next 15 days.

Featured

Is Data Annotation Dying?

Voxel51 co-founder and University of Michigan professor of robotics Jason Corso recently put up an article on data annotation being dead, which sparked discussions on LinkedIn and X alike.

One may have assumed that the wave of generative AI would make data annotation jobs even more abundant. But that is the exact same reason why these jobs are slowly becoming obsolete.

Industry leaders specialising in computer vision solutions stress that despite advancements in AI, meticulously curated, high-quality image-annotated datasets remain essential. These datasets are critical as operations scale and diversify, and the notion that untested technologies could disrupt established workflows is not only impractical but potentially harmful.

Moreover, human-created datasets are proving even more relevant in fields beyond computer vision, extending to generative AI and multimodal workflows. There have been several reports about companies such as OpenAI, Amazon, and Google acquiring cheap labour in countries such as India or even Kenya for labelling and annotating data for training AI models.

In India, companies such as NextWealth, Karya, Appen, Scale AI, and Labelbox are creating jobs within the country, specifically in rural areas, for data annotation. When speaking with AIM, NextWealth MD and founder Sridhar Mitta said, “The beauty of GenAI is that it allows people from remote areas to do these tasks.”

So, are these companies are about to slowly die?

Not So Dead After All

Human annotation has played a pivotal role in the AI boom, providing the foundation for supervised machine learning. The process involves manually labelling raw data to train machine learning algorithms in pattern recognition and predictive tasks.

While labour-intensive, this approach ensures the creation of reliable and accurate datasets. It turns out that the need for human-generated datasets is even more crucial now than ever before.

The only disruption possible is with self-supervised learning. An engineering leader in AI, Tilmann Bruckhaus, said, “These techniques reduce the need for manual labelling by using noisy or automatically-generated labels (weak supervision) or enabling models to learn from unlabeled data (self-supervision).”

Corso believes that human annotation will be needed for gold-standard evaluation datasets, which will also be combined with auto-annotation in the future.

This process involves using AI models to automatically label data. While this approach shows promise, its applicability is limited. Auto-annotation is most useful in scenarios where the model’s performance needs to be adapted to new environments or specific tasks. However, for general applications, a reliance on auto-annotation remains impractical.

Adding to all of this is how current AI models are increasingly relying on synthetic data. SkyEngine AI CEO Bartek Włodarczyk said that with synthetic data “one does not have to worry about data labelling as any masks can be instantly created along with data.”

Dangerous Times Ahead?

Though one can clearly say that human annotation will be the gold standard in the future, if companies fail to adapt and thrive with the current boom, many of them will have to face dangerous times ahead. People For AI founder and data labelling director Matthieu Warnier said, “As labelling tasks become more automated, the ones that remain are notably more complex. Selecting the right labelling partner has become even more crucial.”

This was also reflected by Hugging Face co-founder and CSO Thomas Wolf. “It’s much easier to quickly spin and iterate on a pay-by-usage API than to hire and manage annotators. With model performance strongly improving and the privacy guarantee of open models, it will be harder and harder to justify making complex annotation contracts,” said Wolf, further stating that these will be some dangerous times for data annotation companies.

It seems like manual data annotation might take a backseat when it comes to labelling data for AI training with models such as YOLOv8 or Unitlab’s SAM that can annotate almost anything without a need for human intervention.

On the other hand, manual data annotations will remain a premium service, but the numbers are definitely expected to drop. Companies that are utilising workers in different parts of the world to create high-quality datasets will have to cut down on costs soon.

So, the data annotation market might see major shifts when it comes to adapting to the changing landscape. While the size is definitely set to decrease, the manual data annotation companies will be the ones who set the golden standard, making themselves the benchmark for the automated data annotation market.

The post Is Data Annotation Dying? appeared first on Analytics India Magazine.

The Impact of AI on the Tech Industry

AI is ubiquitous and we all know it by reading about some or the other new product or application showing up in our feed. At the core of AI algorithms lies the data, in the form of numbers. So, let’s talk about the impact of AI in our lives through numbers aka the statistics. Around 77% of devices leverage AI. One of the major uses of AI, specifically the emergence of Generative AI in conversing in human-like language has enabled 85% of AI users to use it for creating content.

Extending it to the industry-wide impact, the retail industry is leading with 72% of retailers employing AI, and 63% of the IT and telecom sector using it. Undoubtedly, it brings remarkable benefits. Take, for example, Netflix which saves $1B using machine learning techniques. But, there is another side to such automation and advancements, i.e. replacing the human workforce. Reportedly, intelligent robots can impact 30% of the global workforce by 2030, impacting 375M careers.

While the numbers could vary from different sources, one thing is clear – AI is rapidly changing the way businesses are run and comes with its pros and cons. So, let's analyze the impact of AI on the technology sector.

Code-Assistants

AI algorithms analyze large amounts of data to generate actionable business insights, helping organizations achieve efficiencies through automation and much more.

Code assistance
Image by Author

Talking about efficiencies, think of how AI-powered code assistants can assist developers in software development by suggesting auto-completion, writing boilerplate code, detecting syntax errors, security vulnerabilities, etc. However, like many other uses of AI, it is recommended to leverage them to augment the developers’ productivity; over-reliance on code assistants to write production-grade code must be avoided.

Cyber-Security

Cybersecurity presents numerous challenges including the presence of a large number of vulnerable systems or devices per organization, diverse attack vectors, a shortage of cyber-skilled professionals, and an overwhelming amount of data. A self-learning, AI-based cybersecurity system could tackle these challenges by continuously gathering and analyzing data from enterprise systems, providing intelligence across areas such as IT asset inventory, threat exposure, control effectiveness, breach risk prediction, fast incident response, and explainability of the model.

Cyber-Security
Image by Author

The advantage of such a technology is that it brings scale and speed, surpassing the limitations of human capability. Tasks such as threat detection and incident management demand analyzing real-time data coming from multiple sources, that can be proactively monitored using AI algorithms.

AI-Powered Chatbots

Be it the code-assistants or the use of AI techniques for cybersecurity, these are closer to the enterprises and are not often evident to their customers. So, let’s shift focus to one of the most pervasive uses of AI in enriching the customer experience through generative AI-powered smart assistants.

AI-Powered Chatbots
Image by Author

Customers need a forum to engage and seek a response to their queries and concerns – and virtual assistants are a great way to address this need. It provides them with information at their fingertips, anytime and anywhere.

Traditional vs Gen-AI Assistants

However, such traditional virtual assistants have been around for some time. So, where do generative-AI-based smart assistants uplift the whole experience?

Generative AI
Image by Author

Well, their ability to converse in human-like language allows them to understand the context very well. This contextual learning involves a nuanced understanding of language, curating a well-formed response that elevates customer experience, and cutting down on time to resolve the query. Gen-AI assistants are able to engage in more natural and fluid conversations leading to a more personalized response.

World of Personalization and Opportunities

Personalization
Image by Author

And yes, the future is all about personalization. It spares the cognitive overload on behalf of the users to apply their custom layer that could cater to their needs. That’s a big driver of using AI in that it speaks directly to the specific user and adapts to their characteristics. While the benefits are many, right from predicting customer behavior, recommendations, content creation, and report generation to infusing AI into existing tech products to streamline user workflows and processes, there is another side to this shiny story of AI advancements.

The Other Side: Workforce Disruption

The rapid integration of AI into business processes has raised concerns about job displacements. The use of AI in automating the repetitive and routine tasks traditionally performed by humans. has led to the displacement of jobs in industries such as manufacturing, retail, customer service, and administrative roles. Not all sectors and occupations are impacted uniformly, in fact, there are new jobs created too. Profiles like prompt engineers, AI ethicists, Responsible AI and Compliance Officers, and AI auditors are fast emerging in the AI-driven economy.

Job Displacement
Image by Author

However, the socio-economic implications for the workforce with low-skilled routine work are facing greater challenges in adapting to technological changes. Policymakers, businesses, and other stakeholders must come together to address these challenges and mitigate the negative impacts on workers and society as a whole. This includes organizing training programs, embedding lifelong learning initiatives, and promoting new job creation in emerging industries.

Ethical Considerations

Not just the workers, but also the consumers of AI-generated responses and predictions are concerned too. The goodness of AI algorithms largely depends on the underlying data they are trained on. Such data may include user information or sensitive details, such as PII (Personal Identifiable Information). Recently, the news has surfaced of companies using users’ data to train their models without asking for any explicit consent. Such unauthorized use of personal data breaches users’ privacy rights and is a big concern when it comes to using AI systems.

AI Ethics
Image by Author

On top of it, the black-box nature of algorithms makes it difficult for not just the users but also for the developers to understand the reasoning behind the model response. Hence, organizations must put concerted efforts into building responsible AI that respects data privacy and promotes model transparency. While AI offers promising benefits such as increased efficiency and the ability to innovate novel solutions, it's crucial to address user and workforce concerns responsibly to fully tap into its potential.

Vidhi Chugh is an AI strategist and a digital transformation leader working at the intersection of product, sciences, and engineering to build scalable machine learning systems. She is an award-winning innovation leader, an author, and an international speaker. She is on a mission to democratize machine learning and break the jargon for everyone to be a part of this transformation.

More On This Topic

  • Celebrating Women in Leadership Roles in the Tech Industry
  • The Ultimate Roadmap to Becoming Specialised in The Tech Industry
  • Popular Google Certification for All Areas in the Tech Industry
  • How to Use Data Visualization to Add Impact to Your Work Reports…
  • From Data Analyst to Data Strategist: The Career Path for Making an Impact
  • IMPACT 2022: The Data Observability Summit, on Oct. 25-26

DSC Weekly 25 June 2024

Announcements

  • The current threat landscape is vast, forcing companies to detect and respond to increasingly sophisticated attacks that cause significant damage. With the proliferation of artificial intelligence (AI) and machine learning, experts are concerned threat actors will weaponize evolving technology for devastating attacks on businesses across the world. The supply chain remains a vulnerable target for attack due to the volume of information being transferred and the third parties involved. Cyber warfare is projected to threaten even more organizations, with nation state attacks making up a growing portion of threats. With all these risks circulated in cyber space, how can companies implement a solid detection and response program that keeps up? Register for the free Enabling Effective Threat Detection and Response summit to learn expert detection and response strategies to safeguard your enterprise from cyber threats.
  • In a world where data is both the lifeblood and a potential liability, staying ahead of data protection trends is vital. The exponential growth of data breaches and cyberattacks, coupled with ever-evolving regulations, demands a proactive data protection strategy. Join the upcoming Data Protection Strategies and Trends summit to discover how to safeguard sensitive information, comply with changing laws, and embrace emerging technologies while building trust with customers. Join us to explore encryption, compliance, transparency, and much more to help navigate the new of era data protection.

Top Stories

  • The Short Lifecycle of Tech Careers
    June 23, 2024
    by Vincent Granville
    It is no secret that when you turn 40, finding a new tech job or avoiding layoffs, becomes a challenge. In this article, I discuss how to prepare for this, and what to do when you get older. Somehow, a tech career has some similarities to professional football players: a few select become millionaire at a young age, many will never make money in tech despite their education, and with few exceptions, you are deemed unfit to play in the field when you turn 40.
  • Talking to a CDO? Think Like an Economist
    June 22, 2024
    by Bill Schmarzo
    The Chief Data Officer (CDO) role has evolved significantly over the past five to seven years. It has transitioned from a CIO Mini-me, focused on managing data infrastructure, to a business executive tasked with deriving value from the organization’s data. Senior management now recognizes the potential of data to optimize operations, mitigate risks, generate new revenue streams, and create a more compelling, differentiated customer experience.
  • Computing: Where should AI safety for superintelligence and AGI start?
    June 24, 2024
    by David Stephen
    It is unlikely that AI safety for superintelligence and artificial general intelligence could be achieved directly without a track from current risks. Already, there are minute fractions of predicted existential threats of AGI that provide a map towards preparing for the unknowns, ahead. AI is a dynamic non-living thing.

In-Depth

  • 8 ways that AI has changed data science forever
    June 24, 2024
    by Anas Baig
    In an era where data is worth more than gold, data scientists are on the cutting edge of business development. But now, a new player has joined the data science game, and its name is Artificial Intelligence (AI). Many in this field wonder how AI will influence their industries, roles, and day-to-day operations.
  • How behavioral biometrics reduces online banking fraud
    June 20, 2024
    by Zachary Amos
    Online banking fraud’s severity and frequency continuously increase, causing staggering financial losses each year. Now that artificial intelligence is in hackers’ hands, traditional security measures aren’t cutting it. Can behavioral biometrics protect online bankers like you? What are behavioral biometrics?
  • How to integrate ChatGPT into your app
    June 19, 2024
    by Erika Balla
    In today’s rapidly advancing technological landscape, integrating AI-powered chat solutions into your applications can significantly enhance user experience and engagement. OpenAI’s ChatGPT, known for its conversational abilities and advanced language understanding, is an excellent choice for developers looking to implement intelligent chat features in their apps.
  • Significance of AI in the development of software products
    June 19, 2024
    by Ryan Williamson
    Artificial Intelligence (AI) is emerging as a formidable force, revolutionizing how we conceive, create, and deliver software solutions. As technology advances at an unprecedented pace, the role of AI in this domain has become increasingly significant. It’s no longer just a buzzword; it’s a fundamental tool that promises to reshape the entire software development process.
  • DSC Weekly 18 June 2024
    June 18, 2024
    by Scott Thompson
    Read more of the top articles from the Data Science Central community.