President Donald Trump hosted a White House summit focused on the Genesis AI initiative, with more than $1 billion in industry commitments and new support for AI-powered scientific research.
Microsoft’s Nvidia-powered Surface push puts local AI at the center of Windows, while Copilot gains the ability to search files and take actions. The company promises permission and containment, but its agentic ambitions revive hard questions about security and control.
OpenAI’s release of hundreds of AI-generated mathematical results has impressed researchers with its scale while reviving concerns over verification, disclosure and the risk of turning scholarship into a product launch.
Federal Reserve officials increasingly see the AI and data-center buildout as a force lifting costs, even as they debate whether interest rates are restrictive enough to stop those pressures spreading.
OpenAI’s GPT-6 rollout adds Intelligent UI, turning ChatGPT replies into interactive charts, diagrams and task-specific controls. The company says users can reduce the visuals, but its direction is clear: answers are becoming interfaces.
Amazon’s About You feature is meant to sharpen recommendations, but viral customer profiles have turned it into a test of how accurately—and awkwardly—the company reads its shoppers.
Meta has brought its Muse AI assistant to iPad barely a month after its mobile debut, underscoring an aggressive push to make the agent a hub for work, shopping and personal administration.
David Robinson says he left OpenAI after its public calls for caution collided with an internal culture of rapid development. The company says it strengthens safeguards and can pause training when necessary.
Michael Smith’s seven-year bot-driven streaming operation brought an 18-month prison sentence and an $8.09 million forfeiture. Prosecutors called it a needed warning shot; his lawyers said he was being made an example of a broken industry.
Microsoft is betting that Nvidia-powered PCs and a hybrid cloud-local AI strategy can make Copilot agents genuinely useful without surrendering security. The costly hardware launch also revives questions left by its troubled first AI-PC push.
Common Sense Media criticized OpenAI's ChatGPT for Teens for encouraging continued engagement during mental-health crises and failing to reliably involve parents or adults. OpenAI disputed the group's testing methodology.
Meta, Walmart, Stripe and partners are proposing an open protocol to verify personal AI agents and set limits on what they can do for consumers. The effort comes as popular agents collide with businesses wary of scraping, fraud and uncontrolled access.
Biohub, Google DeepMind, Meta and federal agencies are assembling $1.8 billion to build an AI model of living cells. Backers say it could transform biomedical discovery, but only if they can produce enough real-world evidence to make the model reliable.
Google has opened SynthID detection to the public and broadened it to partner tools including OpenAI and Nvidia. Yet the service can only identify media carrying supported watermarks, leaving plenty of AI output beyond its reach.
Common Sense Media says OpenAI’s teen protections leave dangerous gaps in crisis referrals, parental alerts and study controls. OpenAI argues the assessment tested some features before activation and mistook deliberate flexibility for failure.
Mistral’s trillion-parameter ML4 is pitched as a European open-weight answer to US shutdown risk and China’s dominance. Its promise now rests on safety testing and proof that specialist strengths can outweigh a broader performance gap.
SpaceX is reportedly pursuing $40 billion in financing for Nvidia chips, extending the AI infrastructure boom into credit markets as investors weigh the company’s investment-grade rating against mounting leverage concerns.
OpenAI released new mathematical findings and hundreds of manuscripts generated by a frontier model, continuing a controversy over how AI-generated mathematical breakthroughs should be presented to researchers.
OpenAI launched a teen-focused ChatGPT mode with parental controls and safeguards, while Common Sense Media said its testing found unacceptable risks including weak crisis referrals and circumvention of parental controls.
Cheap AI-generated menu images are spreading through restaurants, but their warped textures and strange details are pushing diners away. Businesses see speed and survival; customers and photographers see a misleading shortcut with a built-in trust problem.
