AI Rules Watch: What Washington Did on AI Since Labor Day

Washington talked about AI a lot between Labor Day, September 7, 2026, and October 23, 2026, and passed almost nothing binding. For lawyers, the changes worth knowing about are in the courts: the judiciary’s AI task force has issued interim guidance, and the evidence rules committee is still weighing how to treat AI output and deepfakes. Meanwhile the White House issued an executive order on what to call the technology, and Congress traded positions on safety, preemption and privacy without passing anything.

What Are the Federal Courts Telling Judges About AI?

At the Judicial Conference’s September session, Administrative Office Director Judge Robert J. Conrad, Jr. reported that the judiciary’s AI task force has identified more than 60 distinct issues and formed seven subject-matter subgroups. He has issued interim guidance to the courts and said more will follow as issues are resolved. The release says courts have been cautioned not to delegate core judicial functions, including decision-making or case adjudication, to AI, and that all judiciary users have been reminded they are accountable for work done with AI’s help. Ninth Circuit Judge Sidney R. Thomas, who chairs the task force, also briefed the Conference.

“I’m confident that the end result will be a better equipped, stronger, more effective Judiciary,” Conrad said.

Honestly, the biggest change for litigation teams there may have nothing to do with AI. The first component of the replacement for the CM/ECF filing system is on track for release before the end of 2026, with all new district court cases moving to it by the end of 2027. If you run litigation support, ask your e-filing vendor about it now.

Will There Be a Federal Rule on AI Evidence?

Not yet. Proposed Federal Rule of Evidence 707 would apply Rule 702’s reliability standards to AI-generated evidence offered without an expert witness. Its public comment period closed February 16. According to the draft minutes of the committee’s May 2026 meeting, the 59 written comments split three ways: three in unqualified support, 27 supporting it with revisions and 27 opposed. The committee pulled the published version and is studying a revised draft alongside a proposed Rule 901(c) on deepfakes. Both were on the agenda for its October 15 meeting in Boston, whose minutes had not been published as of October 23, 2026.

The May minutes record a clear split between prosecutors and defenders. Elizabeth Shapiro, speaking for the Justice Department, said the department would rather have no Rule 707 and that AI evidence offered without an expert remains an anticipatory problem judges can already manage. Nicole Owens, the Federal Public Defender for Idaho, speaking for the Federal Defenders, said the rule is not premature because defenders are already seeing such evidence, and that AI output doing an expert’s analytical work should face Rule 702 scrutiny. The judiciary’s suggestions docket shows new public submissions on Rules 707 and 901 still arriving through October, along with an October 21 suggestion from Senator Charles Grassley on Civil Rule 26 and third-party litigation funding.

What Does the White House Order Mean for Your Contracts?

Probably very little for now. On September 29, President Trump signed Executive Order 14434, “Inaugurating the Era of Super Intelligence,” which directs agencies, to the extent permitted by law, to use “Super Intelligence” and “SI” in place of “Artificial Intelligence” and “AI” in official correspondence, websites, reports and other non-statutory documents. It does not require changes to existing regulations, contracts or grants, and it ties the new terms to the existing statutory definition of artificial intelligence in 15 U.S.C. 9401(3).

The part lawyers should note is the next step. The order gives the President’s science adviser 60 days, until November 28, 2026, to propose legislative language for a federal definition, including whether it should supersede the statutory definition. A change in statute, unlike a style rule, could reach the AI definitions in contracts and regulations.

The same day, the White House hosted technology executives and released a “Joint Commitment on Frontier Responsibilities,” which Roll Call reported states that “every company is responsible for developing its own technology safely” and that “over time, it may make sense to codify these steps into laws or regulations.”

Where Does Congress Stand on AI Safety and Privacy?

Also on September 29, Senator Mark Warner asked for unanimous consent to pass a bill creating an AI Safety Board within the Commerce Department with enforceable standards for frontier developers. Senator Ted Cruz objected, according to Roll Call. Warner said, “I’ve not talked to a single member on either side of the aisle that doesn’t think we need to move.” Cruz said the bill would give too much authority to the board and the Commerce Department, said he is working on a catastrophic risk bill with Majority Leader John Thune and Senator Amy Klobuchar, and said federal standards “must also address the burgeoning patchwork of state laws.” Speaker Mike Johnson said Congress would “keep steady hands at the wheel on this” without committing to legislation.

On privacy, the House Republican SECURE Data Act (H.R. 8413) would set a national privacy standard enforced by the FTC and state attorneys general. It was not among the 16 bills on the Energy and Commerce Committee’s September 16 markup. The two sides set out their positions at a June hearing. Subcommittee Chairman Gus Bilirakis, in his opening statement, said the bill would create “a uniform national standard that promotes innovation, economic growth, and regulatory certainty.” Ranking Member Frank Pallone said its “sweeping preemption” would “eliminate hard won privacy protections that millions of Americans currently enjoy,” according to his prepared statement.

If your firm or legal department handles client data across states, the question that matters is whether a federal privacy law would set a floor or a ceiling, and as of October 23, 2026 it remains open.

FAQ

How many AI issues has the judiciary’s task force identified?

More than 60 distinct issues, with seven subject-matter subgroups, the judiciary reported on September 17, 2026.

What does the White House order require?

Within 60 days of Executive Order 14434 of September 29, 2026, that is, by November 28, 2026, the President’s science adviser must propose legislative language for a federal definition of the technology.

Did Congress pass an AI law this fall?

No. As Roll Call reported on September 29, 2026, lawmakers traded positions without passing anything, and none had passed as of October 23, 2026.

Were the AI evidence rules adopted?

No. The committee kept working on Rule 707 and Rule 901(c) at its October 15, 2026 meeting, per its agenda book.

*Updated Oct 23 to refresh the story and add an FAQ.*