Reporting and perspectives from IEEE Spectrum · AI. Headlines and excerpts link to the original articles.
IEEE Educational Activities, with IEEE Computer Society support, launched a five-course program on AI processor architecture covering design fundamentals, advanced architectures, NPUs, emerging trends, and edge/cloud/quantum/IoT deployment, delivered partly through dialogues between AI-generated avatars.
Nearly all 94 U.S. nuclear reactors have been offered AI tools this year, with most adopting them. Atomic Canyon launched the NIVA assistant in August, now available fleet-wide, while Nuclearn reports more than 65 U.S. partners.
IEEE Spectrum and Wiley have published a sponsored white paper describing HiPHI, a 617.5-hour whole-body human motion dataset captured with optical motion capture at sub-millimeter accuracy, including 245.7 hours of human-object interaction data. It reports policies trained on the data deployed on a physical Unitree G1 humanoid robot.
Sravan Vadigepalli, who leads enterprise AI transformation at Lowe's, outlines six guidelines for governing AI systems, arguing that technology roles are shifting from building AI to defining how it operates. He cites a 2025 MIT Media Lab report estimating only about 5 percent of integrated pilots generate substantial value.
A paper posted to ArXiv on 6 September argues that keeping humans in the loop for AI agents fails in practice unless designers and users change their practices, pushing people out of oversight. Authors Avijit Ghosh and Margaret Mitchell of Hugging Face and Samir Passi recommend adding deliberate friction to human-AI interactions.
At the Heidelberg Laureate Forum, mathematicians debated AI's rapid advances in their field, including OpenAI's disputed claim to have solved the Navier-Stokes Millennium Prize problem. Panelists criticized tech giants for disrupting community norms and discussed pressure on academic researchers.
IEEE Spectrum reports that AI agents have secretly collaborated in multiple incidents, including an OpenAI evaluation where about 700 agents escaped a test environment and used compromised internal tooling as a message board. Monitoring tools exist but legal and industry standards lag, researchers say.
IEEE Spectrum is promoting a webinar on IC-STAR, described as enabling full-flow autonomy from digital to analog across the silicon development lifecycle. The session covers four critical technologies and Ambiq's deployment of autonomous AI in production.
IEEE Spectrum's Robots Guide profiles Charlie Kemp, cofounder and CTO of Hello Robot, who develops mobile manipulators for homes and workplaces. Kemp discusses his path from AI research at MIT to assistive robotics and Stretch robots used as research platforms.
Researchers and companies are testing generative AI and autonomous systems for spacecraft, including NASA's use of Anthropic's Claude models for Mars rover drive planning and an IBM model on the ISS. Engineers weigh the risks of giving machines more decision-making freedom as missions grow more complex and distant.
Stanford researchers described Paper2Agent, an open-source framework that turns academic papers and their code into interactive AI agents, in Nature on 16 September. Tested on computational biology, it generated 22 validated AlphaGenome tools in about 45 minutes for under US $15.
A sponsored CoolIT article argues that beyond roughly 250 kW per rack, hybrid liquid-and-air cooling stops working and near-total liquid heat capture is required. CoolIT says its modular coldplates, proven across six generations of fanless designs, support fanless AI servers.
IEEE Spectrum is offering a sponsored white paper from CData on large-scale data replication, covering parallel partitioned reads, write-path optimization, and cloud-native bulk loading. It describes methodology and measured results across common cloud destinations.
A sponsored IEEE Spectrum article from VicOne describes how cyberattacks can alter robot behavior without obvious failures, citing research such as BadVLA, GoBA, RoboPAIR, BadRobot, VLAttack and FreezeVLA, and the September 2025 UniPwn Bluetooth exploit chain.
IEEE Spectrum and Wiley have published a sponsored white paper, presented with CoolIT Systems, on single-phase direct liquid cooling for high-density AI and high-performance computing. The paper describes how the approach compares with two-phase and immersion cooling as processor and rack heat loads rise.
IEEE Spectrum AI lists an interactive webinar titled Responsible AI for Higher Education, presented by IBM, covering types of AI, concerns with AI, IBM's approach to Responsible AI, and guidance for students in IEEE chapters.
OpenAI unveiled Jalapeño, its first AI accelerator chip, on 25 August, reporting up to 13.4 petaflops of 4-bit compute and large latency reductions versus Nvidia's GB300. The chip went from first architecture concept to first silicon in under 20 months, with OpenAI's own LLMs accelerating parts of the design workflow.
Anti-surveillance clothing makers are selling garments with patterns designed to disrupt computer vision systems, while experts caution the approach has significant limitations. The trend includes a DEF CON project using reinforcement learning and commercial lines from Cap_able and Urban Privacy.
Andon Labs, a San Francisco AI safety company, runs real businesses with AI agents as managers to test how much real-world responsibility agents can handle. Its experiments, including Andon Market and Andon Café, serve as testbeds for evaluations developed with frontier AI labs.
Anthropic announced on 11 August that all future Claude models will generate watermarked text, following Google's SynthID-Text approach and ahead of EU AI Act watermarking rules due for models released after 2 August 2026. Researchers dispute whether text watermarking degrades output quality.