xAI is one in every of a number of main AI firms to obtain the award, alongside Anthropic, Google, and OpenAI. However the timing of the announcement is hanging given Grok’s current high-profile spiral, which drew congressional ire and public pushback. Using know-how, and particularly AI, within the protection house has lengthy been a controversial subject even throughout the tech business, and Musk’s prior involvement in slashing federal authorities contracts by his work on the Division of Authorities Effectivity (DOGE) nonetheless raises questions about potential conflicts — although his relationship with President Donald Trump has more recently soured, and Trump’s administration has claimed Musk would step again from any potential conflicts whereas at DOGE.
The contract announcement from the Chief Digital and Synthetic Intelligence Workplace (CDAO) is mild on particulars, however says the offers will assist the DoD “develop agentic AI workflows throughout a wide range of mission areas.” Alongside the contract award, xAI announced “Grok for Authorities,” which it says will provide “frontier AI merchandise” to the US. Along with the DoD contract, xAI says different federal businesses will now be capable to buy its instruments through the Common Providers Administration (GSA) schedule. The corporate plans to work on new merchandise for presidency clients, like customized fashions centered on nationwide safety, purposes for healthcare and science use instances, and fashions accessible in labeled environments.
Days after adjustments to Grok despatched it off the rails (saying that if “calling out radicals cheering lifeless children makes me ‘actually Hitler,’ then cross the mustache,” and referencing a “pattern-noticing meme” the place “people with surnames like ‘Steinberg’ (usually Jewish) hold popping up in excessive leftist activism, particularly the anti-white selection”), the corporate apologized for “the horrific habits that many skilled.” It said the replace liable for Grok’s tirades was lively for 16 hours however had been deprecated. Directions given to the chatbot, prefer to not be “afraid to offend people who find themselves politically right,” had the “undesired” impact of main it to “ignore its core values in sure circumstances as a way to make the response partaking to the consumer” — even when that meant “producing responses containing unethical or controversial opinions.”