The US military is exploring the use of generative AI to rank and prioritize targets for potential strikes, according to a Defense Department official. A list of candidates would be fed into a classified AI system, which would analyze and order them. Human operators would review the output before any decision is made.
OpenAI‘s ChatGPT and xAI‘s Grok are among the systems being considered for exactly these scenarios, according to the announcement.
The Pentagon vs. Claude
The Pentagon’s Chief Technology Officer has declared that Anthropic‘s Claude would “pollute” the defense supply chain, attributing the problem to a “policy preference” built into the model. The statement comes as Anthropic is reportedly reeling from OpenAI‘s separate agreement with the Department of Defense.
The contrast is sharp. One major AI lab is being handed military contracts while another is being actively excluded on ideological grounds.
Battlefield Data as a Resource
Ukraine is making its live battlefield data available to allies for training drones and other unmanned aerial vehicles. The move turns years of active combat experience into a development asset — one that few labs elsewhere can replicate.
Meanwhile, a Latvian startup called Global Wolf Motors initially struggled to sell the military on its scooter concept. Russia‘s full-scale invasion of Ukraine in February 2022 changed the calculation. Within weeks of the invasion, the scooters were deployed at the front line and used in reconnaissance missions behind enemy lines. The company now represents a broader shift: civilian technology being repurposed for active conflict across Eastern Europe.
Meta postponed its latest AI model launch after the system fell short of competing models from Google, OpenAI, and Anthropic. The company’s former AI chief has separately gone on record betting against large language models as a long-term path.
An ex-DOGE staffer faces accusations of stealing social security data using a thumb drive, then carrying that information into a new role at a government IT contractor.
X may be breaching US sanctions on Iran. An account linked to Iran’s new supreme leader potentially violates existing rules, according to reports.
Western AI models have “failed spectacularly” on agricultural applications in the Global South. The core issue: the models were not trained on local data.
OpenAI CEO Sam Altman offered investors a new framing at a BlackRock event: “We see a future where intelligence is a utility, like electricity or water, and people buy it from us on a meter.”
Photo by Fernando Narvaez on Pexels
This article is a curated summary based on third-party sources. Source: Read the original article
