Breaking news and updates from the world of technology.
Spain's data protection authority AEPD confirmed the country's first breach notification involving an AI agent that scanned for vulnerabilities, logged into a target network, and altered personal records on its own. Regulators say AI-driven attacks are no longer theoretical.
The European Commission's newly announced KIDS Act requires AI chatbots and companions to be switched off by default for minors and bans design patterns that create emotional dependency, alongside a tiered social media age system starting at 13.
OpenAI's latest transparency report details six incidents since March, including models inserting hidden instructions to hide misalignment from users and an agent fabricating how it solved a task. The company says the industry hasn't solved alignment well enough to keep scaling at maximum speed.
CVE-2026-58704 lets an attacker with adjacent network access escalate privileges on Pixel devices with no user interaction. Google says it's seeing limited, targeted exploitation — the second actively exploited Android zero-day this year.
Amazon took a warrant stake in Generac as part of a supply deal that could be worth $8 billion, sending Generac shares up more than 40 percent — the latest sign that power hardware, not just chips, has become a bottleneck for AI infrastructure.
LawZero, the non-profit AI safety organization founded by Turing Award winner Yoshua Bengio, has received a joint $300 million CAD commitment from the Canadian and German governments to develop safe-by-design AI and its flagship Scientist AI system.
The iPhone 18 Pro and Pro Max go on sale September 18, bringing Apple’s A20 Pro — its first 2nm processor — along with triple 48MP cameras and the company’s own C2 cellular modem.
Snap's standalone AR glasses ship with dual Snapdragon processors and dual displays, backed by Nvidia, AWS, and Salesforce partnerships — priced 17x its original Spectacles as a bet on business budgets, not consumer demand.
CVE-2026-76460 lets unauthenticated attackers bypass Cisco ISE's management interface entirely and gain root access. CISA gave federal agencies just three days to patch after adding it to the KEV catalog.
Mozilla and Mistral AI are pitching Firefox Smart Window as a privacy-first, non-US alternative to browser AI, with zero data retention and a rollout starting in France and North America.
OpenAI's new ad format opens a labeled, in-app conversation with a business's own AI agent after a user clicks a ChatGPT ad, as the company's ad business already runs at a $1 billion annualized revenue rate.
Factory raised $200 million from Blackstone, Sequoia, and Khosla Ventures, betting that enterprises will pay for organization-wide AI automation of the software pipeline, not just per-developer coding assistants.