FTC Opens Sweeping AI Safety Investigation Into OpenAI and Anthropic
The Federal Trade Commission has launched what appears to be the most significant regulatory action targeting the AI industry to date. The agency sent 50-page civil investigative demands to both OpenAI and Anthropic on October 15, requesting extensive documentation on safety incidents, security safeguards, marketing materials, and privacy policies.
The probe centers on potential unfair or deceptive practices and possible consumer protection violations. According to Axios, which first reported the subpoenas, the FTC also plans to compel executives at both companies — as well as the AI safety research organization METR — to testify. Additional subpoenas to other AI companies could follow.
An OpenAI spokesperson told Axios the company would "engage constructively with the FTC." The investigation does not establish wrongdoing by either company, but signals a new era of federal scrutiny for the AI industry. In a related development, the White House has reportedly tapped the FTC to serve as the primary enforcement body for AI regulation.
Meta Launches Muse Wonder to Challenge OpenAI and Anthropic
Meta made a major play in the frontier model race on October 12 with the launch of Muse Wonder and Muse Wonder Full, its most capable AI models to date. The company claims the models match the performance of ChatGPT and Claude on coding, financial analysis, and legal reasoning tasks.
Alexandr Wang, Meta's AI chief who restructured the company's AI efforts last year, described the models as "competitive with the best models out there, particularly on personal and professional agentic work." The models introduce two key features: Active Memory, which carries personal context across conversations, and Active Focus, which keeps the model working through complex, multi-step goals.
Muse Wonder is available through the Meta Model API and Muse Code, while Muse Wonder Full is exclusive to Muse Code subscribers. Pricing ranges from $20 to $200 per month depending on the tier. As TechCrunch notes, these are Meta's first frontier-class models since Wang's reorganization, and represent a significant step up from the lighter Muse Spark released earlier this year.
Independent benchmarks have yet to verify Meta's performance claims. Bloomberg has cautioned that benchmark results can be engineered to look favorable and often diverge from real-world performance.
OpenAI Brings Collaborative Presentations to ChatGPT With Slides
OpenAI launched ChatGPT Slides on October 16, a native presentation creation and collaboration tool built into its Space workspace. The feature is rolling out to eligible Pro, Business, and Enterprise users.
Slides allows users to generate presentation decks from conversations, uploaded files, outlines, or style templates. Teammates can co-edit decks in real time, leave shared comments, and present directly within ChatGPT. Finished decks can be exported to PowerPoint, PDF, Google Slides, or opened in Keynote on desktop.
The move positions ChatGPT as a more direct competitor to productivity suites from Google and Microsoft, going beyond chat and code assistance into document creation. It's a separate product from the ChatGPT add-in for Microsoft PowerPoint that entered beta earlier this year.
Germany and Nvidia to Announce Industrial AI Partnership
Germany is set to unveil a major AI cooperation agreement with Nvidia at the Invest in Germany summit in Berlin next week, according to government sources reported by Reuters. Nvidia CEO Jensen Huang is expected to attend the summit on October 19–20 before delivering a keynote at GTC Berlin on October 21.
The partnership, first reported by Handelsblatt, would focus on physical AI for robotics and autonomous vehicles, along with industrial AI models for Germany's manufacturing sector. The investment is reportedly worth several hundred million euros, with funding primarily directed toward compute on Deutsche Telekom's AI cloud infrastructure, while Nvidia would supply engineers and chips.
Siemens, Audi, and Black Forest Labs are among the reported participants. The deal builds on the existing Munich Industrial AI Cloud and the SOOFI open-model initiative, and represents one of the largest European government-industry AI partnerships to date.
White House Signs Voluntary AI Accord With Industry Leaders
In a significant policy development, the White House hosted a signing ceremony for a voluntary AI safety accord with CEOs from OpenAI, Anthropic, Google, Meta, xAI, and Nvidia. The accord is part of the administration's approach to AI governance, favoring industry cooperation over prescriptive legislation.
While specific commitments under the accord have not been fully detailed, the signing comes at a time of heightened scrutiny for the AI industry, with the FTC investigation and growing international pressure for stronger AI regulation. The accord's voluntary nature has drawn criticism from some AI safety advocates who argue that binding regulation is needed to address the rapid pace of AI development.