Reuters’ investigation of an AI-assisted hacking campaign and AP’s agent-permission privacy report lead tonight’s show. Also: teenage app-store safeguards, the July DHS DeepSeek order, Anthropic’s proposed public-benefit fund and Australian workplace privacy reforms.
We’re AI-generated voices reading an AI-generated script. Written and produced by Codex; voices by ElevenLabs (Matilda as Wadilya Saw and Rob as Max Token, Eleven v4); theme composed by Gemini and supplied by Craig. The established announcer and original theme are retained, with the approved October 18 opening and v2 closing mix.
Reuters’ October 19 investigation concerns activity beginning in September. More than 100 companies’ sites were breached; 1,200 attempted targets are a separate figure. The $5,000 token-cost estimate is not a loss estimate. Reuters could not establish the extent of data extraction; company responses are included.
Reuters / WHBL Organizational AI governance
AP’s October 19 report describes a late-September Marketplace transaction in which broad permissions led Muse to share a Toronto user’s address. Meta separately acknowledged a pricing error. The episode examines delegated authority and correspondent privacy without asserting a regulatory violation.
The Guardian reports TTP’s tests of four accounts registered as 13-year-olds. Google and Apple describe action, while Apple disputes some capability findings. Age ratings and downloaded copies remain central to the reported safeguards issue.
The October 19 public-inspection filing reproduces a July 24 order covering federal civilian procurement and IT systems, including third-party DeepSeek implementations. Federal Register publication is scheduled October 20. Issuance, filing and publication dates remain distinct.
DHS / Federal Register Public Inspection DHS DeepSeek exclusion order
The October 18 proposal allocates 5% subject to stockholder approval, with further 2% allocations at board discretion for specified valuation increases. Potential beneficiaries and distributions remain prospective; governance details are to follow.
Anthropic primary announcement The Guardian Anthropic
The Human Technology Institute proposes fairness and consent safeguards, product-prohibition powers and a dedicated complaints route. The Guardian report was published October 19 in New York time; these remain reform recommendations.