A new bipartisan 269-page AI bill proposes federal preemption of state AI laws, requiring top developers to disclose safety risks while allocating $300 million for enforcement, but critics warn it would strip states of accountability powers and turn existing protections into a ceiling rather than a floor.
Anthropic warns that AI recursive self-improvement could outpace human readiness while simultaneously raising $65 billion and filing for an IPO, exposing a stark contradiction between its safety concerns and aggressive financial expansion — all while Claude already writes over 80% of the company's own code.
OpenAI breaks with the Trump administration by pushing for mandatory AI safety evaluations and civilian oversight bodies, while major tech rivals including Anthropic, IBM, and Amazon back the White House's voluntary framework — though analysts suspect OpenAI's harder stance may be as much about optics as genuine safety concerns.
Anthropic grants the EU's cybersecurity agency ENISA exclusive first access to its powerful AI tool Mythos, making ENISA the first EU body to join Project Glasswing, an initiative allowing select organizations to test the tool ahead of its public release.
OpenAI releases a framework warning that flawed AI testing methods can dramatically skew results, revealing that harness changes alone can cut a model's performance score nearly in half, while calling for full transparency in third-party evaluations.
A sweeping new AI safety benchmark from TELUS Digital exposes alarming security gaps across 34 AI models, finding some comply with harmful requests over 90% of the time across 620,000 attack simulations, with experts warning that most organizations remain dangerously unaware of which vulnerabilities threaten them.
Google DeepMind CEO Demis Hassabis shortens his AGI timeline to 2029-2030, citing rapid AI agent acceleration, as tech leaders race to predict when artificial general intelligence will reshape humanity.
Cisco research exposes a critical AI safety flaw, revealing that every major frontier AI model from OpenAI, Anthropic, Google, Amazon, and xAI fails multi-turn attack tests, with vulnerability rates rising up to 71% after just five conversational turns — exposing dangerous gaps in current industry safety benchmarks.
As public trust in AI collapses — with 70% of Americans opposing local data center construction and youth enthusiasm plummeting 14 points — the FBI and DHS are now actively surveilling a rising wave of anti-AI extremism fueled by job displacement fears and coercive workplace mandates.
Pope Leo XIV issues a landmark 44,000-word encyclical demanding a ban on lethal AI weapons, limits on Big Tech power, and protections for children online, framing AI development as a defining moral crossroads between profit-driven dominance and human dignity — with even Anthropic co-founder Christopher Olah attending the Vatican presentation …