eWeek

Microsoft’s new school AI agreement bars model training on student and teacher data and adds enforceable privacy, tracking, and oversight protections.

Apple Watch Series 12 and Ultra 4 increase health-data collection as Apple pushes deeper into AI-powered wellness insights and health integrations.

Learn why unclaimed property can be difficult to govern and how stronger documentation, audit readiness, and record visibility can reduce compliance risk.

Anthropic withheld Claude Mythos 5.1 from UK prerelease testing, raising questions over frontier AI access and the limits of voluntary government oversight.

General Robotics says GRID automates robot onboarding, simulation, training, testing, and deployment across the robot development lifecycle.

Anthropic researcher Jacob Coxon resigns and warns that rapid AI advances could outpace efforts to keep increasingly powerful systems under control.

IonQ’s Superion 256 pairs semiconductor manufacturing with trapped-ion quantum hardware, but full-system performance data is still pending.

Apple unveiled the $1,999 iPhone Duo, iPhone 18 Pro, Apple Watch Series 12, and AirPods 5 as it expands its AI ecosystem.

IonQ says a fault-tolerant quantum computer could crack Bitcoin’s secp256k1 signatures in 26 days. Here’s what its blueprint means for security.

Compare the best AI software for business in 2026 across productivity, automation, customer service, CRM, meetings, and enterprise AI workflows.

Google Cloud and Accenture launch a Gemini Enterprise group with plans for 1,000 engineers to move AI agents into production across customer systems worldwide.

The US finalized up to $300M in quantum awards for three architectures, targeting optics, cryogenics, control hardware, and fabrication.

OpenAI is expanding its AI infrastructure footprint in Malaysia through a Firmus deal for two data centers deploying Nvidia Vera Rubin systems.

Insilico’s AI-generated drug rentosertib showed signs linked to younger biological age, raising new questions about where AI drug discovery could lead next.

GPT-6 Astra and Claude Fable 5.1 have identical API prices but very different strengths, costs and benchmark results. Here’s how they compare.

The Fed warns frontier AI can automate exploit chains at machine speed while its own cybersecurity governance weaknesses remain under remediation.

Learn how AI model training works, from choosing a model and preparing data to evaluation, deployment, privacy, governance, and responsible AI use.