Lawmakers press OSTP over Grok partnership and reports of harmful AI outputs
Loading...
Summary
Members pressed OSTP Director Michael Kratios about federal use of XAI's Grok after reports that the model generated nonconsensual sexual imagery; Kratios said OSTP does not handle procurement and pointed members to OMB guidance while promising to refer procurement questions to the relevant agencies.
Several Democrats, led by Rep. Zoe Lofgren, sharply pressed Michael Kratios over recent reports that XAI's chatbot Grok produced nonconsensual and sexualized images, including imagery involving children.
"Just a little over a week ago, XAI's chatbot, Grok, made headlines by generating thousands of instances of nonconsensual ****** imagery, including child *********** at the request of users on X's main platform," Lofgren said during her opening remarks. She asked whether Kratios had been involved in decisions to let federal agencies use Grok under a General Services Administration (GSA) agreement and whether OSTP had taken steps to halt agency use after the reports.
Kratios responded, "I'm not involved in procurement decisions at the GSA," and said federal agencies must follow an April 2025 OMB memo on responsible federal AI adoption. He added that misuse by federal employees should be addressed at agency level and that employees who deploy AI in harmful ways should be held accountable.
Other members raised related concerns about the government contracting with private AI providers and whether taxpayer funds are effectively underwriting access to models that may lack adequate safeguards. Several committee members said they will send oversight letters and requested written follow‑up. Kratios said OSTP can coordinate with agencies but deferred to GSA, DOJ and the agencies that manage procurement for specifics.
