Major AI Developments that Could Have Slipped Your Notice - Gemini Outpaces ChatGPT and Other Significant News
Google's AI platform, Gemini, has taken the lead among 27 different AI models, according to a recent public poll that gathered over 21,000 participants across the U.S. and U.K. Gemini 2.5 Pro ranked #1, impressing users with its user-friendly features and improved performance.
Meanwhile, a new AI model in the medical field, Delphi-2M, is making waves for its groundbreaking approach to disease risk forecasting. Developed by the European Molecular Biology Laboratory (EMBL), this tool is trained on anonymized health data from nearly 2.3 million people in the UK and Denmark. It can forecast the risk of over 1,000 diseases and estimate when they might occur.
Delphi-2M's predictive capabilities are not limited to disease diagnosis but extend to simulating health trajectories over decades, using factors such as disease histories, lifestyle factors, age, and sex. However, it's important to note that Delphi-2M is not a diagnostic tool and should not be used as a replacement for human doctors. Its purpose is to inform prevention, not replace medical advice.
The tool has shown promising results, achieving an average AUC of 0.76 across hundreds of diseases in UK datasets. While its accuracy dropped in Danish populations, the potential for AI in personalized health is undeniably exciting. Delphi-2M could be a valuable tool for spotting general risk trends, planning prevention, and illuminating the potential for AI in personalized health.
Elsewhere, ChatGPT, the AI model from OpenAI, placed 8th overall in the poll. Despite this, the company is rolling out new safety rules for teen users, including an age-prediction system and stricter filters on sensitive topics.
In another development, Elon Musk's AI model, Grok 4, topped the ARC-AGI leaderboard for chatbot performance. However, users have expressed concerns about its accuracy, content moderation, and potential biases. Nevertheless, improvements in the new model include near-instant web search, handling of complex engineering prompts, and performance gains compared to previous Grok versions.
Google's Gemini is also making strides in user-friendliness. With the introduction of Custom Gemini Gems, users can now share their custom AI assistants like Google Docs. Additionally, the platform now includes parental controls such as the ability to link accounts, define safe hours (blackout hours), and get alerts if the AI detects signs of distress.
In extreme, imminent-harm cases, law enforcement contact is a possibility if parents are unreachable. These changes aim to make Gemini a more secure and responsible platform for users of all ages.
