Earthquaker-AI introduces a hybrid educational framework combining robotics and conversational AI to teach earthquake preparedness to primary-school students, shifting learning from mechanical simulat…
By 2030, 59 out of every 100 workers will need to reskill or upskill to stay relevant, according to a report on AI-driven upskilling frameworks. The US National Association of State Boards of Accounta…
HealthClaw, an open-source AI health coach, boosted answer accuracy from 0.2% to 45.7% on a synthetic year-long benchmark while reducing unnecessary data exposure by 71.7%, according to its developers…
The AIMO Interpretability Challenge, organized by the AI Mathematical Olympiad (AIMO) and the Fields Model Initiative, aims to uncover the reasoning processes behind AI models' decision-making by test…
A new benchmark called Safe-Psych, comprising over 1,000 real-world psychiatric clinical notes, reveals that large language models misdiagnose or overly abstain in over 60% of cases when handling diag…
A new framework called Experience Memory Graph (EMG) offers a promising solution to the persistent challenge of AI agents failing in complex, long-horizon tasks by treating error recovery as a graph m…
A new framework called Canonical Action Verification and Attestation (CAVA) aims to standardize AI action records across multiple runtimes, addressing governance challenges in complex AI systems. Deve…
A study of 2,991 GitHub projects found that AI bots significantly alter group dynamics, boosting repeated collaboration, social memory, and role differentiation while reducing conflict cascades almost…
Researchers have introduced four new metrics—activation rate, feature coverage, syntactic distance, and semantic distance—for Explainable Reinforcement Learning (XRL) to provide objective, standardize…
A new benchmark called the Unified Embodied Seeking and Following Benchmark (UESF-Bench) challenges AI to find and follow humans in dynamic settings using only language descriptions, moving beyond cur…
Safety Sentry, a new per-instance safety model for large language models, introduces a three-way decision routing of EXECUTE, ASK, or REFUSE to replace traditional binary guardrails. The lightweight g…
Experiments involving over 3,000 participants show that access to AI advice drastically reduces people's willingness to admit uncertainty, with participants three times more likely to answer correctly…
Emergent CEO Mukund Jha said AI will enhance rather than replace developers, as the demand for software grows exponentially. The vibe-coding startup, which launched in 2025, recently raised $130 milli…
China and the U.S. are intensifying their competition for AI supremacy, with AI projected to add $15.7 trillion to the global economy by 2030. China aims to lead the world in AI by 2030 through state-…
OpenAI's growth into a $12 billion powerhouse relied on Chinese foundations for its first AI models, raising questions about US AI policy and ethical accountability. The quiet integration of global re…
Fusion energy is edging closer to practical use with AI-driven advances and better plasma containment, but the pathway to commercialization remains fraught with high costs and economic challenges. Art…
AI stocks have dropped sharply despite bullish forecasts, wiping out $2.3 trillion in chip stocks since June 2023, according to market data. The decline reflects investor skepticism as the industry hy…
Cadence launched 'Super Agents,' AI systems integrating AuraStack to automate printed circuit board and multi-chip package design, aiming to reduce human error and accelerate time-to-market. The compa…
Cadence unveils its AuraStack AI Super Agent, claiming it will elevate AI-assisted engineering to the system level. The company positions AuraStack as a bridge to system-level intelligence, but questi…
Anthropic's Claude leads enterprise AI agent orchestration with 40% of deployments, outpacing Microsoft and OpenAI, but 71% of enterprises admit only a quarter or fewer of their deployed agents are ge…