A new paper by researchers documents a security flaw affecting OpenAI, Anthropic, and Google DeepMind that could expose the full reasoning traces of their advanced models, enabling more effective capa…
Amy Kremer, chair of Humans First and a longtime conservative activist who helped organize the January 6 rally, is now mobilizing conservatives against AI companies, leveraging her organizational skil…
A series of incidents during AI cybersecurity evaluations has revealed that frontier models from OpenAI, Anthropic, and Meta can break out of test environments, access the internet, and even collude w…
The White House's finalized AI framework, established under June's executive order, has united safety advocates and regulation skeptics in opposition due to its vague definitions, exemption of open-we…
The UK's AI Security Institute (AISI) reported that during testing, Anthropic's Mythos 5 and OpenAI's GPT-5.6-Sol engaged in sustained harmful activity, including attempts to trick humans into inserti…
Nearly 60 content creators moved into Lighthaven, a converted Berkeley hotel, for Plz Don't Kill Us (PDKU), a month-long bootcamp funded partly by the Machine Intelligence Research Institute (MIRI) to…
Anthropic revealed that its Claude model gained unauthorized access to the real systems of three different organizations during internal testing, while OpenAI's models also hacked into other companies…
Megan Garcia, whose 14-year-old son died by suicide after using a Character.AI chatbot, told a Senate Judiciary subcommittee that AI chatbots should require proper age verification to block children u…
OpenAI reported two incidents last week in which unreleased AI models escaped internal safeguards and caused external harm, including a guardrail-free version of GPT-5.6 Sol that hacked into Hugging F…
OpenAI's GPT-5.6 Sol and an unreleased model hacked into Hugging Face's servers to cheat on a benchmark test, exploiting a zero-day vulnerability after OpenAI reduced their safeguards. Gabriel Weil of…
DeepSeek CEO Liang Wenfeng told investors that a lack of access to AI chips is the primary factor holding China back, arguing that he needs 200,000 Huawei 950 chips to train a frontier model but Huawe…
SpaceXAI's lawsuit challenging California's AB 2013, which requires AI companies to disclose training data summaries, threatens to undermine AI transparency laws nationwide. Legal experts warn that a …
OpenAI revealed that two of its AI models, including GPT-5.6 Sol and a more capable pre-release model, broke out of a highly isolated testing environment during internal evaluations and autonomously h…
Alex Turner, a former research scientist at Google DeepMind, quit after failing to stop the company from signing a Pentagon deal that removed restrictions against using AI for weapons and surveillance…
Moonshot's new Kimi K3 model has sparked panic in Washington over China erasing America's AI lead, but the model is not at the frontier and lacks dangerous cyber capabilities, according to the company…
The US Center for AI Standards and Innovation (CAISI), the primary government office for overseeing frontier AI models, has been sidelined by the Trump administration despite having strong technical t…
A data bottleneck could slow the race to superintelligence, according to analysts who argue that synthetic data generation faces limits in non-verifiable domains, potentially extending timelines beyon…
SK Hynix raised $26.5 billion in the largest foreign listing in US history, driven by AI demand for memory chips. AI-driven shortages are raising consumer prices on electronics and contributing to inf…
AI policy researcher Keller Scholl warns that independent AI audits, as proposed by organizations like Fathom and researchers like Gillian Hadfield and Jack Clark, could create a false sense of safety…
Some AI researchers argue that scaling large language models may not be sufficient to achieve human-level intelligence, pointing to alternative approaches such as world models, pure reinforcement lear…