Every story in this issue with a short summary, so you can decide what to read before leaving the site. Summaries are written by AI (Claude) only from each publisher's own description and checked by Jev for anything not in that description. They are not full translations. Follow "Read the original" for the full story.
Two lawmakers introduced frontier AI model safety bills in Harrisburg yesterday, the first introduced after the OpenAI / Hugging Face attacks. The bills contain strong pre-deployment requirements that would strengthen standards set earlier this year in Illinois and California.
Read the original →Every AI system that can use an organisation’s data and systems needs a named human owner, a defined approval boundary and a record of what it did. These must be in place before the system acts.
Read the original →A phishing message no longer needs to persuade the employee who owns an inbox, because it can target the software reading that inbox on their behalf. This is the central warning of IRONSCALES' new report, From Deepfakes to Double Agents.
Read the original →Ambassador Cha Ji-hoon will co-chair preparations for the United Nations' 2027 Global Dialogue on AI Governance.
Read the original →The Holy See has warned about the modernization and expansion of nuclear arsenals, coupled with increasingly prominent rhetoric.
Read the original →Innovare Limited, a management consulting and technology solutions provider, has announced a governance and management systems implementation partnership.
Read the original →OpenAI CEO Sam Altman laid out three paths to safe AI in an internal presentation: an international watchdog modeled after the IAEA, a merger of the major labs into one entity, or a single company so powerful it could shut down everyone else's compute. He called the last option "terrible."
Read the original →As AI grows more powerful, two communities are clashing over how to keep it under control and who gets to decide whether it is safe.
Read the original →Digital and technology ministers from the Group of Seven (G7) nations held an online meeting on the 20th. They agreed to strengthen international cooperation.
Read the original →According to the SANS Institute, roles mitigating AI deepfakes and overseeing post-quantum encryption rollouts are on the horizon.
Read the original →NIST, CISA and NSA list ten practices for preparing for, detecting, responding to, recovering from, and reporting rogue AI agent incidents.
Read the original →Gray Swan's new Arena includes $300,000 in prize money for developing tactics to jailbreak models from leading AI companies.
Read the original →Intel unveiled its D-Series AI inference GPU using up to 480GB of low-power LPDDR5X DRAM instead of HBM. The product targets capacity-focused enterprise buyers.
Read the original →Many questions around how it will work remain.
Read the original →The analysis argues that artificial intelligence can bring justice closer to those who have always been far removed from it.
Read the original →On 30 September 2026, the Tokyo District Court considered a dispute between a Japanese voice actor and the video-sharing platform TikTok over videos that used a voice allegedly generated or processed using AI to resemble his own. The Court indicated that a person's voice, like their likeness, may serve as a symbol of their personality and may be protected by publicity rights in certain circumstances. The Court did not determine whether the use of the disputed voice actually infringed any rights.
Read the original →A Haywood County resident has filed a lawsuit against Tennessee education policymakers. The suit alleges that the subcommittee responsible for crafting AI policies for K-12 classrooms has been working in secret, violating the state's Sunshine Laws.
Read the original →