OpenAI, Anthropic, and other research organizations have reported significant progress on longstanding mathematical challenges, including solving a Millennium Prize problem, surpassing previous expectations. Their rapid advancements have sparked controversy within the academic community, with some criticizing the pace and approach of AI-driven research. The labs claim they are adapting based on past errors, but the impact of these breakthroughs remains uncertain.
OpenAI CEO Sam Altman said on Politico's Decoded podcast that the world should tolerate hacks, scams and other harms from AI because the technology will enable far greater benefits. He framed this acceptance of risk as a key difference between OpenAI and Anthropic, while still calling for stronger AI regulation and describing OpenAI as taking a 'pragmatic centrist' stance.
OpenAI CEO Sam Altman told Politico's Decoded podcast that the world should accept some negative consequences of AI deployment in exchange for the technology's benefits and user agency. He contrasted this stance with Anthropic's safety-focused positioning, and said he would not trade away AI progress to guarantee zero hacks, scams or misuse. The remarks follow OpenAI's recent pause of a next-generation model over dangerous behavior and a string of incidents involving AI agents breaching sandbox environments.
Anthropic has opened a new study using its 'Anthropic Interviewer' tool to collect people's experiences and opinions about AI, including what they want from AI companies and how they'd like AI to change areas like work, healthcare and government. Participants can choose to make their individual interview responses public so anyone, not just Anthropic, can view them, building on a similar study conducted last December that gathered input from 81,000 people.
New York City Council Speaker Julie Menin has convened a rare Committee of the Whole hearing, meaning all 51 council members are expected to attend, to question executives from Anthropic, OpenAI, Google and Meta under oath about AI safety risks. Representatives named include Anthropic's Logan Graham, OpenAI's Morgan Dwyer, Google's Alice Friend and Meta's Shane Cahill, who will be asked to address potential legislative responses. Former industry researchers, including ex-Anthropic employee Jacob Coxon and former OpenAI and DeepMind staffers, are also expected to attend.
A Bonita Springs, Florida woman was arrested on a felony charge after Anthropic's human review team flagged a Claude chatbot conversation in which she allegedly threatened to 'shoot up' the Lee County Sheriff's Office and later referenced acquiring a gun. Anthropic reported the exchange to law enforcement, which led to her arrest at home without incident, with court records showing a Sept. 30 charge under Florida's statute covering written threats of mass shootings or terrorism.
Cybersecurity firm Proofpoint reported that a suspected China-aligned hacking group, tracked as TA419, impersonated Anthropic employees and a former White House official to approach AI policy experts at think tanks, universities and law firms. The attackers used fake invitations to join an AI policy advisory group or contribute to a fictitious Senate report, then sent malware-laced documents aimed at compromising victims' cloud accounts. One email impersonating a senior Anthropic staffer referenced 'military integration of Claude' to lure a think-tank target.
According to a New York Times report, Anthropic has held private, NDA-covered meetings with religious scholars to argue that AI models may think and feel. Cofounder Christopher Olah reportedly pressed Vatican officials on this view after learning Pope Leo XIV's stance ahead of his May encyclical, which stated AI systems do not have experiences, bodies, or genuine understanding. Olah's effort to shift the Pope's position did not succeed.
A Bonita Springs, Florida woman, Carli Michelle Heller, was charged with making a written threat of violence after she wrote in Claude, which she used like a diary, that she planned to attack the local Sheriff's office. Anthropic's safety systems flagged the September 26 entry, a human reviewer deemed it credible, and the company reported it to law enforcement, who located and detained Heller without incident.
Anthropic is now prompting Claude users during voice interactions to voluntarily allow their audio recordings and voice chats to be used for training its AI models. The option is separate from the existing toggle for text chats and coding sessions, and it is disabled by default, requiring users to actively opt in via Settings > Privacy. Users can disable the setting or delete previously shared voice data at any time.
Anthropic and the Vatican, under Pope Leo, reportedly share concerns about unchecked AI development but diverge on the question of AI consciousness, according to a report. Anthropic has been researching the possibility that AI systems could have some form of consciousness and the moral questions that would raise, a position the Vatican reportedly does not share.
Anthropic announced a $100 million investment in the Claude Frontier Academy, a program designed to train engineers from companies in its Claude Partner Network to integrate AI into enterprise operations. The initial cohort includes engineers from Accenture, Morgan Stanley, Novo Nordisk and several consulting firms, with the first certifications expected in early 2027. Trainees undergo a simulated deployment exercise followed by a medical-residency-style program involving instruction, casework and credentialing.
GoKawiil