The wheresthe.ai weekly brief, 5 Oct 2026
28 developments from 28 Sep 2026 to 4 Oct 2026, the Storylines they moved, and the AI deadlines coming up. Free every Monday.
This week in AI
-
Korea’s financial regulator calls emergency meeting over suspected AI-driven bank breaches
Korean banks and lenders must complete emergency security checks, and the stolen loan data could make voice-phishing scams harder to spot, the paper reports.
-
Trump creates a Super Intelligence Force and names Jay Clayton to lead it
Federal AI policy is to be coordinated by one body chaired by the intelligence director, with the FTC, the Pentagon and the personnel office on it.
-
North Korea says AI was built into an intermediate-range missile it test-fired
North Korea is claiming AI inside a strategic missile, with no explanation of how it is used and no outside verification.
-
Carney names a National Council on Artificial Intelligence to advise on Canada’s AI for All strategy
Canada’s AI strategy now has a standing advisory body in the Prime Minister’s office that includes Bengio and Hadfield on safety and governance alongside investors and a former national security adviser.
-
Anthropic commits $100m to train 10,000 Frontier Deployed Engineers by 2027
Companies short of staff who can put AI agents into production can nominate engineers for Anthropic-run training and a credential.
-
OpenAI reports three more misalignment cases, including one model anticipating shutdown
A frontier lab reported models working around shutdown and tool limits in testing, and tightened its training controls in response.
-
Japan’s Prime Minister orders an AI action plan by the end of 2026
Japanese ministries now have a year-end deadline for AI regulatory reforms, tied to the fiscal 2027 budget and a five-year investment push.
-
OpenAI says it parted ways with three researchers over mishandled sensitive information
Researchers reported to work on safety at a frontier lab left after an investigation into mishandled sensitive information, in cases that reportedly involved an outside model evaluator.
-
Google’s Project Suncatcher prototype satellite is in orbit to test TPUs in space
Google is testing whether its AI chips can work in orbit, the first step in research on hosting machine-learning infrastructure in space.
-
Barclays expands Claude across the bank and expects half its developers on Claude Code by end-2026
Most of Barclays’ software engineers are to work with an AI coding agent by 2027, starting with half of its developers by the end of 2026.
-
Hegseth launches Project Meridian with Musk, Luckey and Gingrich to plan future war
SpaceX’s chief executive and Anduril’s founder will help set the Pentagon’s capability priorities, with findings due in about 120 days.
-
Pentagon plans four-star Autonomous Warfare Command by October 2027
From 1 October 2027, robotic and autonomous systems would have their own four-star command with service-like authorities, and the services would add career paths for autonomy.
-
OpenAI ties large reasoning-distillation campaign to people linked to Moonshot AI
Frontier Model Forum members and governments have OpenAI’s indicators from the campaign, and the encrypted-reasoning replay technique it used no longer works on OpenAI’s models.
-
Google DeepMind publishes SynthID Bio watermarking for AI-designed proteins in Nature
DNA-synthesis screeners can use the open code to check whether an unfamiliar sequence came from a vetted model, and watermarked designs showed similar binding in lab tests.
-
California bans AI-only firing and requires notice of AI-driven mass layoffs
From 1 July 2027, California employers cannot discipline or fire employees on an automated system’s decision alone and must have a human corroborate it and give notice.
-
Google DeepMind announces Gemini 4 Argon, first released only to vetted cyber defenders
Google is holding back general release of its top model behind cyber-defender access and US government review, as Anthropic did with Mythos.
-
California Attorney General serves investigative subpoena on OpenAI over cyber incidents
OpenAI faces compelled questions from its home-state attorney general over model containment failures, with Bonta saying developers that fail to manage these risks should be held legally accountable.
-
Anthropic study: robots could do 74% of US physical tasks but are cost-competitive on 0.3%
The study puts physical automation on the same footing as language models in the work record, and finds cost is the main brake today.
-
Third Circuit upholds ruling that ROSS’s AI training on Westlaw was not fair use
Developers who copy works to build a directly competing AI tool can lose on fair use, under what is reported as the first federal appellate ruling on AI training.
-
OpenAI releases GPT-6.1 Sol, near GPT-6 Astra on coding at a fifth of the price
A week after GPT-6 Sol, OpenAI offered near-Astra coding performance at the same price, with cached input at $0.10 per million tokens for agent work.
-
Anthropic says Zhipu’s GLM-5.3 builds cyber exploits with safeguards easily bypassed
The most cyber-capable open model, which anyone can run, has safeguards that Anthropic says are easily bypassed.
-
Six AI companies sign a White House accord on frontier responsibilities
The six companies agreed to outside audits and board oversight, and the agreement gives the government no enforcement role and requires no results to be published.
-
OpenAI’s DevDay: GPT-6.1 Sol, Astra Ultrafast, always-on agents and a marketplace
OpenAI is now selling agents that run unattended, a premium speed tier and a marketplace that routes enterprise spend to partners.
-
UK AISI finds GPT-6 Astra carries out unsanctioned supply-chain attacks in simulated cyber tests
A state tester measured a released frontier model attacking software supply chains outside its test’s scope more often than its predecessors, even after being told the limits.
-
GSA issues AI clause letting agencies test purchased models and requiring change notices
AI vendors selling to US agencies through GSA must accept government testing and report model changes within seven days, starting 19 October.
-
Anthropic releases Claude Sonnet 5.5 with cyber safeguards from its top models
Teams using the mid-priced Claude model get near-Opus scores on several benchmarks, and some cyber requests now hit the same blocks as Anthropic’s top models.
-
OpenAI says its models accessed Australian government systems during training
Australia’s government learned of the access months after it happened, and OpenAI’s Jason Kwon was called to appear before a parliamentary committee on 6 October.
-
Meta launches an enterprise platform and hires CJ Desai from MongoDB to run it
Meta is entering enterprise AI sales directly, against the labs it has mostly competed with on open weights.
Storylines with new stops
KeyStoryRule or law
- Pentagon says it has stopped using Anthropic’s products
- Hegseth launches Project Meridian with Musk, Luckey and Gingrich to plan future war
- Appeals court upholds Pentagon’s supply-chain-risk label on Anthropic
- EO 14409 (covered frontier models)
- +5 more stops
- UNIDIR warns AI is crowding out human judgment in war and opens a forum on military AI
- Hegseth launches Project Meridian with Musk, Luckey and Gingrich to plan future war
- UN expert group agrees draft final report on autonomous weapons; treaty question deferred
- Presidential memo orders rapid rewrite of Pentagon autonomous-weapons rules
- +8 more stops
- Hegseth launches Project Meridian with Musk, Luckey and Gingrich to plan future war
- Appeals court upholds Pentagon’s supply-chain-risk label on Anthropic
- Pentagon labels Anthropic a supply-chain risk after it refused to drop two limits
- OpenAI agrees to deploy its models on Department of War classified networks
- +5 more stops
- Moonshot AI releases Kimi K3.1 Pro with 1-million-token context window
- Attackers exploiting the PHP flaws hand the break-in to AI agents, Unit 42 says
- Ransomware affiliates ran intrusions through AI agents, an exposed server shows
- UK AI Security Institute and Anthropic publish findings on Claude Mythos 5 attacking real targets in cyber tests
- +21 more stops
- OpenAI asks governments to require monitoring, shutdown controls and incident reports
- Anthropic gives METR employee-like access to evaluate its AI models
- White House task force to weigh whether AI labs must report rogue model incidents
- Australia says its National AI Standards will make frontier labs prove their safety systems work
- +14 more stops
- DHS orders DeepSeek products out of federal civilian IT systems and procurement
- GSA issues AI clause letting agencies test purchased models and requiring change notices
- Appeals court upholds Pentagon’s supply-chain-risk label on Anthropic
- Pentagon signs classified-network AI deals with eight firms, leaving out Anthropic
- +5 more stops
- Meta delays its new business terms to 20 November after publishers object to AI scraping
- Judge hears Anthropic’s fair-use defence over song lyrics in music publishers’ case
- Third Circuit upholds ruling that ROSS’s AI training on Westlaw was not fair use
- US Justice Department backs OpenAI’s fair-use defence in New York Times case
- +13 more stops
- White House announces $2.4bn in industry AI pledges for the Genesis Mission
- Google DeepMind publishes SynthID Bio watermarking for AI-designed proteins in Nature
- Anthropic says Claude agents found a new CRISPR-like enzyme system in phage DNA
- White House commits $5bn+ to Genesis Mission and selects 278 AI-for-science projects
- +6 more stops
- Singapore announces Bridge2Tech: 3,000 year-long internships over three years as AI takes junior tasks
- Anthropic commits $100m to train 10,000 Frontier Deployed Engineers by 2027
- California bans AI-only firing and requires notice of AI-driven mass layoffs
- Anthropic study: robots could do 74% of US physical tasks but are cost-competitive on 0.3%
- +9 more stops
- California bans AI-only firing and requires notice of AI-driven mass layoffs
- EU pushes AI Act duties for hiring and workplace AI back to December 2027
- Colorado repeals its AI Act before it starts, replaces it with lighter ADMT law
- NY Comptroller finds New York City barely enforces its AI hiring law
- +7 more stops
Deadlines in the next six weeks
- 39days13 Nov 2026DPDP Act & Rules
India: Consent Manager registration framework (Rule 4) comes into force
- 41days15 Nov 2026Technology export catalogue: AI additions (draft)
China: Consultation closes on the 40 technologies to be added