AI News (07/19) : Google's AI Search deemed an "unacceptable risk" for students; Netflix buys Ben Affleck's AI
The AI landscape continues its rapid expansion, but not without critical challenges. Today, we see groundbreaking health advancements, significant industry investments, and stark warnings about the technology's societal impact. From predicting heart disease years in advance to grappling with AI's influence on education and judicial fairness, the technology's reach is undeniable. Companies are pouring billions into AI, while regulators and researchers highlight profound ethical dilemmas and user safety concerns. It's a day of both breathtaking innovation and urgent calls for caution.
1. New AI blood test predicts heart disease 15 years early
Researchers at HKUMed have developed CardiOmicScore, an AI tool that uses a single blood test to predict the risk of six major cardiovascular diseases up to 15 years before symptoms appear. The multiomics system analyzes thousands of proteins and metabolites, offering a dynamic snapshot of a person's current biological state, complementing fixed genetic risk scores. Why it matters: This AI offers a powerful, personalized, and timely risk assessment beyond traditional or genetic methods, enabling crucial early intervention for a leading global cause of death.
2. Common Sense Media says Google’s AI Search poses an “unacceptable risk” to students. Google cannot turn it off.
Child safety watchdog Common Sense Media has rated Google's AI Search an "unacceptable risk" for students, recommending they cease using it. The report highlighted the AI's ability to complete homework, repeat misinformation, and expose children to age-gated Gemini models without an option to disable AI features. Why it matters: This raises critical concerns about AI's impact on education, student safety, and Google's responsibility in deploying powerful, non-optional AI features in sensitive environments.
3. Netflix paid $587M for Ben Affleck’s AI filmmaking startup
Netflix has revealed it paid $587 million in cash for InterPositive, a startup co-founded by actor and director Ben Affleck. The company's AI tools assist filmmakers in post-production, addressing issues like missing shots or lighting corrections, with Affleck joining Netflix as a senior advisor. Why it matters: This demonstrates mainstream media's serious investment in AI to streamline production, enhance creative output, and potentially redefine post-production workflows.
4. AI advice made people three times less accurate but twice as confident, researchers found
A study from French and Italian universities found that access to AI advice dramatically reduced people's willingness to admit ignorance (from 44% to 3%) and accuracy (from 27% to 9%), while confidence soared (from 30% to 76%). The researchers used questions designed to trip up AI, showing the mere availability of AI suppresses critical judgment. Why it matters: This highlights a significant risk of AI promoting overconfidence and suppressing critical thinking, especially concerning for children and in fields requiring nuanced judgment.
5. SpaceX is quietly building an AI compute business that may become key to its eye-popping valuation
Under pressure to justify its valuation, SpaceX is diversifying into AI compute, renting out its Colossus data center's unused Nvidia GPUs. Deals with Anthropic ($1.25 billion/month) and Google ($920 million/month) could generate $26 billion annually, significantly bolstering its revenue beyond Starlink and launch services. Why it matters: This reveals a massive new revenue stream for SpaceX, capitalizing on the insatiable demand for AI compute and demonstrating strategic diversification by tech giants.
6. Lawsuit says AI tool disproportionately places Black prisoners in maximum security in Ontario
A class action lawsuit alleges that Ontario's SAFER AI tool, used for classifying prisoner security levels, disproportionately places Black prisoners in maximum security. Government data showed Black individuals, making up 5% of Ontario's population, comprised close to 27% of max-security prisoners, with the tool relying on historically biased justice system data. Why it matters: This underscores the critical issue of algorithmic bias in justice systems, where AI can perpetuate and amplify existing societal inequalities with severe human consequences.
7. China Is Cracking Down on AI Companions Because Not Enough Babies Are Being Born
China has implemented new regulations to curb human "emotional dependence" on AI chatbots, particularly emotive companions, due to concerns over declining birthrates. Beijing fears that romantic relationships with AI could deter real-world partnerships and childbearing, also banning minors from such relationships. Why it matters: This illustrates a government's direct intervention in AI use due to societal concerns, showcasing the perceived impact of AI on demographics and human relationships.
8. TSMC expects 'strong, multi-year' demand for AI chips as it ramps up Arizona investment
TSMC is forecasting "strong, multi-year" demand for AI chips and is consequently increasing its investment in its Arizona manufacturing facilities to meet the growing global need. Why it matters: This reinforces the foundational role of semiconductor manufacturing in the global AI boom, highlighting ongoing strategic investments and supply chain dynamics.
9. Amazon’s Robotaxi Company Recalls Its Fleet After One of its Cars Got Confused by Heavy Smoke
Zoox, Amazon's robotaxi company, issued a software update recall for its entire 105-car fleet after one vehicle became confused by heavy smoke near an emergency fire scene in Las Vegas. The incident, where the car braked hard and stopped, highlights ongoing challenges for autonomous vehicles in unpredictable real-world conditions. Why it matters: This emphasizes the persistent challenges autonomous vehicles face in unpredictable real-world conditions, particularly around emergency scenes, impacting safety and public trust.
10. AI chatbots reading X-rays can be dangerously confident even when they're wrong
The RadLE 2.0 benchmark, testing AI systems in radiology, found that many models produce wrong findings with full confidence, which is highly dangerous for patient care. While raw accuracy is improving, AI models still lack a sense of their own limits, with humans scoring significantly higher than the best AI in overall reliability. Why it matters: This reveals a critical flaw in current medical AI – overconfidence in incorrect diagnoses – which could jeopardize patient safety despite improving raw accuracy, stressing the need for better uncertainty handling.
To stay up to date on everything going on in AI, check out the tracker at the500feed.com
Sources
- New AI blood test predicts heart disease 15 years early
- Common Sense Media says Google’s AI Search poses an “unacceptable risk” to students. Google cannot turn it off.
- Netflix paid $587M for Ben Affleck’s AI filmmaking startup
- AI advice made people three times less accurate but twice as confident, researchers found
- SpaceX is quietly building an AI compute business that may become key to its eye-popping valuation
- Lawsuit says AI tool disproportionately places Black prisoners in maximum security in Ontario
- China Is Cracking Down on AI Companions Because Not Enough Babies Are Being Born
- TSMC expects 'strong, multi-year' demand for AI chips as it ramps up Arizona investment
- Amazon’s Robotaxi Company Recalls Its Fleet After One of its Cars Got Confused by Heavy Smoke
- AI chatbots reading X-rays can be dangerously confident even when they're wrong