United is changing its booking rules to limit AI agents from buying flights through its website and travel sellers. The airline says agents could expose customer information, make unwanted purchases or run afoul of aviation rules that require a person to complete certain steps.
OpenAI's quarterly revenue rose about 50% to $10 billion in Q3, helped by a doubling of its business with companies. It still trailed Anthropic's $17 billion in revenue and reported a $10.9 billion adjusted operating loss.
Google has introduced the Gemini Decisions API, which uses text, images, audio, video, and PDFs to make quick choices from a set of options. It provides confidence scores for tasks such as choosing tools and checking policies, and is available through Google AI Studio.
Helsing has unveiled Clarity, an AI system that uses military data to suggest targets and suitable weapons. The company says a person must approve each target before it is passed to a weapons system, and every decision will record who was responsible.
Nvidia shares fell about 8% in after-hours trading Wednesday amid reports that its board may consider removing CEO Jensen Huang. The reports follow disclosures about Huang’s personal relationships, and Nvidia is reportedly looking for possible successors.
CoreWeave’s upcoming earnings report will test whether it can ease concerns about the debt it has taken on to expand its AI data centers. Its stock and bonds have fallen as investors question how it will fund more construction and repay its loans.
Security firm Claroc found that nearly 95% of the Grok-built apps it checked were accessible without a login. Some exposed patient information, unencrypted passwords, and confidential sales records from SpaceXSI, which makes Grok.
Meta’s longtime infrastructure leader, Santosh Janardhan, is stepping down and expects to leave next year after helping with the transition. Dave Brown, a former Amazon executive, will take over as Meta expands its data centers for AI.
Big Walk developer House House disclosed that it used an AI chatbot to help solve a bug, with a small amount of the resulting code included in the game. The admission has intensified a debate among indie game award judges, some of whom refuse to vote for games made with generative AI.
New York has launched a state-run program to independently test advanced AI models, with researchers at Binghamton University leading the work. The center will study risks to inform public decisions, but it can examine only information companies make public.