The year 2024 was all about artificial intelligence. From finding mention in almost every industry to breakthroughs that seemed straight out of science fiction, AI’s progress was infectious.
Among the standout achievements were an AI foundation model, named Chef, capable of accurately detecting multiple types of cancer; AMD stepping up to rival NVIDIA and Intel with its new AI chipset; and a wave of exploration into AI agents by tech giants and startups alike.
Yet, amidst all this innovation, one question keeps raising its head: Are all AI advancements truly necessary? Maybe slapping an AI label on products and services doesn’t automatically make them revolutionary, or even relevant to humanity’s immediate needs.
As we approach 2025, let’s look into some AI-powered inventions that may not be solving the problems of today but are certainly raising eyebrows.
Wearable AI Can’t Replace Smartphones Yet
In a rush to bring AI closer to us in our daily lives, gadgets like the Limitless AI Pendant and the WUZPR Ring were introduced in 2024. These devices promised to harness the potential of AI and enable users to interact seamlessly.
Then there’s Elon Musk’s Neuralink, which aims to merge the human brain with technology through implants. It’s designed to facilitate communication with machines and even other humans. The vision is audacious but also raises questions about its necessity and the ethical implications of such tech.
When the Humane Ai Pin was unveiled, it was seen as a game-changer. Packed with features like contextual insights, virtual assistance, and personalised recommendations, it sounded like the ultimate wearable tech. However, reality didn’t live up to the hype. Early users expressed disappointment, with many feeling that the product overpromised and underdelivered.
As of now, AI is just being integrated into smartphones as the makers see a chance to move towards the new technology. According to the latest update, Qualcomm and MediaTek have introduced smartphone chipsets that enable the processing power required for AI applications.
In 2023, Samsung unveiled its groundbreaking generative AI model, Samsung Gauss, marking a significant leap forward in artificial intelligence technology. Meanwhile, Google introduced its AI-powered Google Workspace suite, showcasing the increasing integration of AI into everyday tools and services.
AI is Now your Boss’ Best Friend
AI is no more your buddy; it has now taken on the role of a hyper-attentive colleague silently watching you at work. The new ChatGPT desktop app, powered by OpenAI, holds the ability to monitor screens in real-time, this AI assistant is ready to jump in during your moments of crisis.
Microsoft isn’t far behind. CEO Satya Nadella unveiled Copilot Vision, a supercharged, rebranded version of their Recall feature. Forget basic keyword searches; this AI does a deep dive into your digital history, reconstructing past moments.
Meanwhile, Anthropic’s Claude 3.5 pushes the boundaries further, simulating human-like tasks on your computer screen with its new “computer use” feature.
What does all this mean? In a micromanaged corporate world, these AI tools might replace your manager, who keeps a tab on your work. While these features are marketed as productivity boosters, they add another layer to the surveillance-heavy culture of today’s workplaces.
Robotaxi is Not Yet available for a Ride
Elon Musk has finally promised to bring Tesla’s robotaxis to the streets of California and Texas next year, claiming autonomous vehicles will give people their time back. But let’s push the brakes. Experts predict that instead of easing traffic, a swarm of self-driving cars could actually worsen congestion on the already jam-packed roads.
Also, Tesla may be late to the party. Alphabet’s Waymo has been running commercial robotaxi services in San Francisco and Phoenix since 2021. Meanwhile, GM’s Cruise is teaming up with Uber to launch autonomous rides in select cities by 2025.
And what about India? While the country struggles with the low adoption of electric vehicles, the idea of automated taxis feels like a distant dream. With its never-ending traffic issue, robotaxis might face a steep uphill drive in terms of acceptance.
AlphaFold 3 is Also an Experiment
In May 2024, Google DeepMind, teaming up with Isomorphic Labs, released AlphaFold 3, an AI model that pushes the boundaries of biomolecular understanding. This latest innovation doesn’t just predict protein structures like its predecessor; it ventures further, deciphering the intricate structures and interactions of all biological molecules, from proteins and DNA to RNA and ligands.
For the first time, an AI system has outperformed traditional physics-based tools in predicting biomolecular structures, an achievement that redefines what’s possible in computational biology.
The implications are monumental, particularly in drug discovery. AlphaFold 3’s precision in predicting how small molecules interact with biological targets could dramatically accelerate the creation of new medications, erasing challenges like costs and timelines.
AlphaFold 3 builds on the revolutionary foundation of AlphaFold 2, which came in 2020 by solving the decades-old protein-folding problem, a challenge critical to biology and medicine. Its accurate prediction of 3D protein structures from amino acid sequences reshaped the landscape of research and therapeutic innovation.
But this innovation came in late. Four years of refinement and experimentation have culminated in AlphaFold 3, proof that game-changing science often requires patience, persistence, and a touch of bold ambition.
OpenAI 01 Does What!
OpenAI recently dropped a chatbot powered by the OpenAI o1 technology. This new model promises advanced reasoning skills, tackling challenges in science, coding, and maths with a level of finesse that its predecessors couldn’t match. It’s the first in a series that’s considered for regular upgrades, setting high expectations for ChatGPT’s future capabilities.
But here’s the twist: OpenAI isn’t rushing to flaunt o1’s maths skills on traditional benchmarks like MATH and GSM8K. Why? They’re calling these benchmarks outdated, claiming that most modern AI models, o1 included, can breeze through them without breaking a sweat.
In OpenAI’s words, “Recent frontier models do so well on MATH2 and GSM8K that these benchmarks are no longer effective at differentiating models.”
What’s Next?
As the buzz around AI continues, it is now the time for AI agents in automation, business, finance, and insurance. These are some of the sectors actively adopting the technology. Research suggests that organisations that scale AI across their finance functions tend to perform better and achieve higher ROI.
So, maybe, as of now, the coming years are for AI agents and nothing more. With the momentum building, it seems like everything else is just background noise.
The post The AI We Don’t Need appeared first on Analytics India Magazine.
