The Elon Musk vs. OpenAI trial opened on April 27 in federal court before Judge Yvonne Gonzalez Rogers. Musk testified for nearly two hours, telling the jury he 'was a fool' for funding what became a startup, and warning about a 'Terminator outcome' from unregulated AGI development. OpenAI counsel William Savitt walked Musk through 2017-era exhibits showing Musk had explored a for-profit OpenAI where he would hold majority equity and control — directly contradicting his 'stole a charity' framing.
Microsoft sits as co-defendant on an aiding-and-abetting claim. The jury's verdict will be advisory; Judge Rogers will make the final call. The case has drawn intense interest as a referendum on whether OpenAI's transition from nonprofit to for-profit violated its founding mission. Day two featured OpenAI CEO Sam Altman's testimony, where he acknowledged tensions with Musk but defended the for-profit pivot as necessary to compete with Google and attract the capital needed for frontier AI development.
The European Union's AI Act entered its most significant enforcement phase, with rules for high-risk AI systems — including those used in critical infrastructure, education, employment, law enforcement, and healthcare — becoming fully enforceable. Companies deploying high-risk systems must maintain detailed technical documentation, implement risk management protocols, and pass third-party conformity assessments before market placement. Open-source model developers releasing general-purpose AI models must now provide training data summaries and energy consumption reports.
The enforcement represents the world's first comprehensive AI regulatory framework in action. Non-compliance carries fines of up to 7% of global annual turnover or €35 million. The EU AI Office has begun conducting investigations, with several unnamed US-based AI companies reportedly under preliminary review. Industry groups have criticized the cost of compliance, but the EU has signaled no intention to delay enforcement.
Google signed a classified AI contract with the Department of Defense on April 28, giving the Pentagon access to its AI for 'any lawful government purpose.' The contract includes language stating the AI 'is not intended for, and should not be used for, domestic mass surveillance or autonomous weapons (including target selection) without appropriate human oversight' — but explicitly does not give Google authority to veto operational decisions. That last clause is a direct response to the Anthropic standoff: the Pentagon designated Anthropic a supply-chain risk after the company refused to remove guardrails.
The contract closes a complicated chapter in Pentagon AI procurement. The DoD signed $200M agreements with Anthropic, OpenAI, and Google in 2025. After Anthropic's blacklisting, the Pentagon signed replacement deals with seven companies including Google, OpenAI, Microsoft, Amazon, Nvidia, SpaceX, and startup Reflection. Google's signing completes the circle — and confirms that the Pentagon's real dispute with Anthropic was about veto power, not guardrails.