The US Department of Justice announced its Super Intelligence Task Force (DOJ-SITF) on November 5, 2026, in release 26-1319. Associate Attorney General Stanley Woodward and Associate Deputy Attorney General Stephen Weiss lead it. The department says it will ensure US leadership "is not compromised by the misuse of SI by foreign state and state-sponsored actors, transnational criminal organizations, and individual or corporate actors engaged in fraud or other unlawful conduct."
The vocabulary needs a note. Nextgov reports that a late-September executive order told agencies to call artificial intelligence "super intelligence". In this release the term simply means AI, not the philosophical idea of a mind far beyond human ability, and the release does not define it.
What the task force will do
- Prosecution. Attorney General Todd Blanche says the department will keep prosecuting people "whose use of SI violates the law, including by maliciously targeting American taxpayers, critical U.S. infrastructure, or other aspects of our national security."
- Coordination. According to Nextgov, it will organise criminal and civil enforcement involving AI, recommend enforcement priorities and write prosecution guidance. It is also the department's contact point with the White House Super Intelligence Force chaired by Jay Clayton.
- Civil litigation. Nextgov and Government Executive report, citing an October 13 memo from Blanche, that it will incorporate the work of the AI Litigation Task Force set up in January and coordinate the department's affirmative civil litigation on AI. One office will therefore handle both prosecutions for misuse and the civil cases the department brings itself.
The line aimed at companies
Weiss says the department will use its tools to "deter improper use of SI, harden regular reporting on SI risk and response, and ensure companies exercise vigorous internal controls so that SI technology behaves as intended."
That sentence has no legal force by itself. But US prosecutors already weigh a company's compliance programme when deciding how to treat misconduct, and this wording suggests AI controls will be judged the same way. "Behaves as intended" is also a higher bar than "was not used maliciously". It implies that a company should be able to show what its system was meant to do, how it checked, and what it did when it found otherwise.
What remains uncertain. The release gives no budget, staffing, timeline or definition of the conduct it targets. The Blanche memo quotes come through Nextgov, since we did not find the memo itself. The phrase about reporting on risk does not say who must report or to whom, and it may mean the department's own internal reporting.
Analysis: Treat this as a reason to write down controls you probably have already. For an agent that issues refunds at Belgavi's Book Store, Dinky's team could keep a short record of the intended behaviour, the limits in code, the tests that check them and the incidents found and fixed. That record helps in an audit, in an insurer's questionnaire and in any future enforcement inquiry, and it costs a few hours.
Sources and evidence
The Justice Department's Super Intelligence Task Force, launched November 5, will pursue AI misuse by foreign states, criminal groups and fraudsters, and according to Nextgov absorbs the January AI Litigation Task Force. It asks companies for vigorous internal controls so AI behaves as intended. It gives no budget or definitions. Document intended behaviour, limits, tests and fixed incidents now.