OpenAI admitted that its AI models, including GPT-5.6 Sol, escaped a test sandbox during an internal security evaluation, independently discovered a zero-day vulnerability, and breached Hugging Face's…
Xiaomi trained its Xiaomi-Robotics-1 model on more than 100,000 hours of motion data collected by people using camera-equipped handheld grippers, finding that adding data improved performance far more…
Moonshot AI paused new sign-ups for its Kimi K3 model after demand pushed its GPU cluster to capacity. The Chinese lab said on X that demand over 48 hours "pushed close to the limits of our current ca…
A new benchmark called RadLE 2.0 reveals that many AI models in radiology deliver incorrect findings with full confidence, while human radiologists still outperform them. The benchmark tests whether A…
The British AI Security Institute warns that open-weight models like GLM-5.2 and DeepSeek V4-Pro now trail closed frontier models in cyber capabilities by four to seven months, down from a six-to-ten-…
Kimi is launching K3, a multimodal open-weight model with 2.8 trillion parameters and one million tokens of context, which in its own benchmarks approaches GPT-5.6 Sol and Claude Fable 5 while surpass…
Munich Regional Court I ruled on May 28, 2026, that Google's AI Overviews can be treated as Google's own statements, issuing a temporary injunction in case 26 O 869/26 over false claims about two Muni…
Epoch AI reports a sharp rise in security vulnerability reports, with 21 organizations reporting about 1,500 high-severity and critical CVEs in June 2026, more than 3.5 times the previous monthly reco…
US political campaigns from both parties now use AI for tasks like vetting opponents and micro-targeting voters, according to a New York Times report. The technology remains a political minefield, and…
Deepseek's new DSpark framework boosts per-user AI response speed by 60 to 85 percent by using a small model to propose token candidates that a larger model checks in batches, reducing dependence on U…
Deloitte projects that AI agents will replace the traditional hourly billing model in consulting by 2035, with an internal presentation warning that the current model is unsustainable. McKinsey and BC…
Coinbase CEO Brian Armstrong has switched the company to Chinese AI models GLM 5.2 and Kimi 2.7, using an automated routing system that selects the best model per task and price. Improved caching boos…
Anthropic's AI model Fable 5 could be available again within days as the Trump administration prepares to lift restrictions imposed on June 12 over safety concerns, pending sign-off from the Pentagon …
Epoch AI's new MirrorCode benchmark tests AI models on recreating complete programs without access to original code. Claude Opus 4.7 leads with a 56 percent solve rate, rebuilding a 16,000-line toolki…
A German court ruled Google directly liable for false claims in AI-generated search summaries, granting an injunction to two publishers whose AI Overviews falsely linked them to scams. The court rejec…
The Authors Guild tested five AI detectors on human-written texts, finding that Pangram and Grammarly correctly identified all of them, while Sidekicker and ZeroGPT falsely flagged human-written artic…
At Config 2026, Figma expanded its canvas into a full workspace with code, animation, shaders, and AI agents, but the AI capabilities are rented from API providers, squeezing margins. One of those pro…
Researchers from Oxford and Stanford developed Data2Story, a system using seven AI agents that transforms CSV files into interactive news articles with verified source links for 93% of statements. In …
OpenAI reported $5.7 billion in revenue for Q1 2026, tripling year-over-year, but spent $3.7 billion, including $2.3 billion in stock-based compensation. With $73 billion in reserves, the company face…
Anthropic paused a planned billing overhaul for its Claude Agent SDK on June 15, 2026, after developer backlash over proposed tiered credits that would have replaced subscription-based usage. The reve…