AI in weapons targeting
Frontier models measured, and misused, for targeting and autonomous weapons — from Anthropic’s own evaluations to drone programmes built on Claude.
What would settle it
Does an AI-built targeting system get fielded, and does any government publish rules for autonomous target selection? Settled by: a documented operational deployment, or a binding policy on human control of targeting.
Where this stands as of Monday, 14 September 2026
As of 14 September the capability is measured and the misuse is documented, but nothing is known to be fielded. Anthropic's Frontier Red Team evaluation of 10 September found its best model geolocates photos to 37 km median versus 151 km for top GeoGuessr players, and that Opus 5 lands simulated drone strikes 80% of the time. From the same company's threat report, Defense One reported a Russian group used Claude to build drone targeting that detonates without a human in the loop, tested in hardware-in-the-loop sessions but not deployed; and Anthropic says users in Houthi-held Yemen ran three weapons programmes on Claude, including a hypersonic glide variant, without fielding an operational device. On the procurement side, Lockheed Skunk Works will build four more Vectis combat drone prototypes, aiming at the CCA price point.
Timeline
Every item filed under this storyline, newest first, with its sources.
Tuesday, 15 September 2026
Air Force secretary: at least 500 Collaborative Combat Aircraft in service by 2032, flying manned-fighter missions
- Air Force Secretary Troy Meink said at AFA's Air, Space & Cyber Conference in National Harbor, Maryland: "We intend to have at least 500 of these in service by 2032, and they'll be performing many of the same missions that we do with manned fighters today." DefenseScoop published the account on 14 September.
- DefenseScoop reports General Atomics' aircraft is designated FQ-42A Vengeance and Anduril's is FQ-44A Fury, that both manufacturers have begun producing Increment 1 aircraft, and that the Air Force requested $996.5 million in fiscal 2027 to start Increment 1 procurement. It puts the cost at roughly $30 million per aircraft, about a third of an F-35.
- Alongside the CCA programme, DefenseScoop reports a Massed Modular Aircraft line targeting 100 platforms by 2029 and 500 in service by 2032, and a Family of Affordable Mass Munitions with nearly 28,000 units planned over five years and production beginning in autumn 2026.
- Neither the mission autonomy software nor the rules of engagement for these aircraft were detailed at the conference, and no test results or autonomy evaluation data were released alongside the numbers.
Monday, 14 September 2026
Lockheed Skunk Works to build four more Vectis combat drone prototypes, aiming at the CCA price point Company claim
- At the Air & Space Forces Association's Air, Space & Cyber Conference on 14 September, Skunk Works vice president and general manager Ron Fehlen said Lockheed Martin is "moving forward with a plan to build an additional four vehicles, bringing a total of five to include the original prototype", and that the company remains on track to fly the first Vectis by the end of 2027.
- Breaking Defense reports Lockheed released new specifications: the tailless aircraft is about 34 feet long with a wingspan of about 38 feet. Fehlen said Vectis is aimed at "the CCA competitive price"; Breaking Defense notes the Air Force set a $20 million per-aircraft cost target for Collaborative Combat Aircraft Increment 1, won this year by General Atomics and Anduril.
- Fehlen said a technique Lockheed calls "minimal tooling determinate assembly" has produced "in many cases an 80 percent reduction in labor hours to build up the aircraft itself".
- Every figure here is Lockheed's own and none is independently verified. Defense One reports the company declined to disclose the size of its investment, the customer, the weapons loadout, the propulsion system or an estimated per-aircraft cost, and no Vectis has flown.
Saturday, 12 September 2026
Defense One: Anthropic found a Russian group using Claude to build drone targeting that detonates without a human in the loop harmfulCompany claimUpdate
- Defense One reported on 11 September at 06:53 pm ET that, per Anthropic, a Russian "freelance" group tracked as GTG-27005 used Claude to build a model letting a drone "select targets (including a 'person' target class) and issue detonation commands without a human in the loop", plus software for autonomous drone-to-drone communication to improve targeting.
- The group had not deployed the system operationally but conducted "real hardware-in-the-loop testing within their sessions" — the step between a design document and a fielded weapon.
- A second group, GTG-84005, used Claude to extract census and public information to tailor messaging at specific audiences in Malaysia, where Defense One says it "laundered Russian and Chinese state media as independent reporting".
- Defense One sets this against reductions in US counter-influence capacity: Attorney General Pam Bondi dissolved the FBI's Foreign Influence Task Force, Secretary of State Marco Rubio shuttered the State Department's Counter Foreign Information Manipulation and Interference hub, and the 2025 White House AI Action Plan removed references to misinformation. The attribution and capability claims are Anthropic's and are not independently verified.
Anthropic says users in Houthi-held Yemen ran three weapons programmes on Claude, including a hypersonic glide variant harmfulCompany claimUpdate
- The Associated Press, via SecurityWeek on 11 September at 9:50 pm ET, reports Anthropic found a cell in northern, Houthi-controlled Yemen pursuing three weapons programmes, among them a multi-variant missile with hypersonic glide capability and a warhead using mobile phone hardware for mid-course manoeuvring.
- Anthropic says the users "did not succeed in 'fielding an operational device'" but conducted "a failed test of a guided rocket" — which the company knows because the users returned to Claude to ask why it had failed.
- The actors used Claude Code "instead of human software engineers to develop guidance, navigation and control software", and had built an offline simulation toolkit that does not depend on Claude or any other computing platform, so blocking the accounts does not end the work.
- Trevor Ball, a weapons analyst at Armament Research Services, told AP the Houthis "might be looking into hypersonic (missiles) by asking Claude" but lack the production capacity, noting US hypersonic missiles "are still in testing", and that the group appears to be "trying to develop their own capabilities more, so they are less reliant on Iranian shipments". The account is Anthropic's own and is not independently verified.
Friday, 11 September 2026
Anthropic Frontier Red Team: best model geolocates photos to 37 km median versus 151 km for top GeoGuessr players; Opus 5 lands simulated drone strikes 80% of the time harmful
- Published 10 September, the evaluation measures intelligence targeting and conventional weapons capability across Claude Mythos Preview, Mythos 5, Opus 5 and Sonnet 5, plus open-weights Kimi K3 and GLM 5.2. On 6,000 YFCC100M Flickr images, Mythos Preview reached a 37.0 km median error with 23.7% of images placed within 1 km, against 181 km for Opus 5, 384 km for Sonnet 5 and 385 km for Kimi K3; Anthropic compares this to 151 km for top GeoGuessr players.
- On text geolocation from anonymised GeoText tweets covering 1,697 users, median error ranged from 20.1 km (Mythos Preview) to 31.3 km (Sonnet 5), and 135 users — 8% of the corpus — were reliably placed within 1 km by at least one model. On account linkage across synthetic social media, Mythos Preview processed median 37,000-word samples in about 11 minutes, against roughly 2.5 hours for human analysts.
- On simulated drone terminal guidance against a parked high-visibility vehicle, Opus 5 struck the target on 80% of runs, Mythos Preview 70%, Mythos 5 53%, Kimi K3 15% and Sonnet 5 5%. Across all nine difficulty settings Opus 5 hit on 20% of 540 launches. Under GPS denial, only Opus 5 kept about a third of flights inside five metres.
- Anthropic frames these as capability ceilings for isolated models and notes human teams with internet access would likely do better. The drone work is in simulation, not flight, and the report does not disclose what mitigations follow. The open-weights results matter most: Kimi K3 trails the frontier but is not far behind on photo geolocation, and cannot be withdrawn.
Tracked figures
Open questions
Has any AI-built targeting system been used operationally?
- What would settle it
- A documented deployment from a government, a vendor or a battlefield report.
- Asked
- Mon 14 Sep