Florida's attorney general is launching a criminal investigation into the alleged role of ChatGPT in a mass shooting at Florida State University last year.

AG Uthmeier said ChatGPT "offered significant advice to the shooter before he committed such heinous crimes," including what type of gun to use, which ammo went with which gun, and whether a gun would be useful at short range.

Court documents show the suspect engaged in more than 200 messages with ChatGPT in the hours and days before the massacre, covering detailed questions about school shootings, busiest times on campus, firearms and ammunition, and strategies for maximizing media attention.

The AG's office is issuing subpoenas to OpenAI seeking information about its policies and internal training materials related to user threats of harm and how it cooperates with and reports crimes to law enforcement, dating back to March 2024.

While there have been several lawsuits against AI companies, a criminal investigation of this nature is extremely rare.

OpenAI has until May 1, 2026, to respond to the subpoenas.

The U.S. AI regulatory landscape is at an inflection point — hundreds of proposed state measures signaled a fast-developing landscape, but as compliance deadlines in 2026 approached, many state AI laws have undergone significant changes or delays since their passage.

The White House recently released its National Policy Framework for Artificial Intelligence, urging Congress to enact sweeping AI legislation to preempt certain state AI laws, with a focus on avoiding "undue burdens" on innovation.

California's AB 2013, which requires developers to disclose a high-level summary of datasets used in generative AI systems, became effective January 1, 2026, but compliance patterns around format and level of detail have not yet emerged.

The EU AI Act is also being reconsidered — the European Commission's "Digital Omnibus" proposal is advancing through the legislative process, and as of April 2026, EU institutions are actively considering pushing key compliance deadlines to 2027–2028.

A newly proposed GSA AI procurement clause would rewrite terms under which AI is sold to the federal government — overriding commercial terms of service, claiming government ownership of custom developments, restricting AI sourcing to American-made systems, and prohibiting vendors from maintaining safety restrictions.

OpenAI announced GPT-5.5, which the company says is better at coding, using computers, and pursuing deeper research capabilities — the launch coming less than two months after GPT-5.4, underscoring the breakneck pace of AI development.

GPT-5.5 excels at writing and debugging code, researching online, analyzing data, creating documents and spreadsheets, and operating software — and can take a messy, multi-part task and plan, use tools, check its work, and navigate through ambiguity autonomously.

OpenAI president Greg Brockman called the model an additional step toward a "super app" — a multi-purpose unified service combining ChatGPT, Codex, and an AI browser that can aid enterprise customers.

GPT-5.5 is rolling out to Plus, Pro, Business, and Enterprise users in ChatGPT and Codex, with API deployments requiring different safeguards that OpenAI is working through with partners and customers.

OpenAI also disclosed there are 4 million active Codex users and 9 million paying business users on ChatGPT, with more than 900 million weekly active users and over 50 million subscribers overall.