Curated AI news from the last 24 hours — models, agents, enterprise, and developer tools.
At its September event, Apple unveiled the foldable iPhone Duo alongside the iPhone 18 Pro and Max and a new Watch Series 12, marking the first major product launch under new CEO John Ternus. AI featured heavily in the lineup, with Siri integrating Google's Gemini model to answer questions using context pulled from a user's texts and email.
cnbc.com ↗Apple introduced a new way for iPhone users to verify that their photos have not been altered or generated by AI, adding content-authenticity signals to help counter the spread of AI-generated images passed off as real photography. The feature was announced alongside the rest of Apple's AI-focused September hardware lineup.
techcrunch.com ↗Microsoft unveiled 'Codename MDASH,' an agentic AI system that brings autonomous security-scanning capabilities to US government agencies. The tool is designed to let AI agents proactively identify and flag cybersecurity vulnerabilities across government IT systems.
microsoft.com ↗Microsoft partnered with teachers' unions AFT and UFT to announce a national AI safety and privacy standard for schools, aimed at protecting students, families, and educators as AI tools become more common in classrooms. The standard sets guidelines for responsible AI use in K-12 education settings.
news.microsoft.com ↗A researcher at Anthropic resigned and publicly warned against pursuing self-improving AI systems, describing the current pace of frontier development as 'gambling with our lives.' The departure adds to a string of safety-focused exits from major AI labs this year.
techcrunch.com ↗Google Cloud and Accenture are forming a joint unit to send engineers into enterprises to help them deploy Google's AI tools, following similar enterprise-implementation pushes from OpenAI, Anthropic, and Microsoft. The deal is aimed at helping Google catch up in the competitive race to convert AI model advances into paid enterprise deployments.
techcrunch.com ↗OpenAI appointed a well-known critic of unchecked AI development, associated with warnings about existential risk from advanced AI systems, to its board of directors. The move adds a safety-focused governance voice to the board as the company continues to race to deploy increasingly capable models.
techcrunch.com ↗Meta is offering users of its Muse Spark 1.3 AI model discounts averaging about 95% in exchange for letting the company use their prompts and the model's outputs to help train future models. The opt-in data-sharing program targets developers and direct users of the model rather than Meta's consumer apps.
techcrunch.com ↗A new report found that AI spending per employee at top companies slumped in August, raising questions about whether the dip reflects normal summer seasonality or an early warning sign for enterprise AI budgets. The data adds to a broader debate about the pace and durability of corporate AI adoption heading into the fall.
techcrunch.com ↗