Policy

Jeffries Staffer Built AI Tool to Track Lawmakers, Draft Talking Points

A House Democratic leadership aide created Caucus Pulse to monitor social media and generate messaging as the party campaigns on AI regulation.

Omega Editorial · November 21, 2026 · 4 min read

Key takeaways

  • A Jeffries staffer built Caucus Pulse, an AI tool that analyzed 114,647 social media posts and drafted messaging for the Democratic leader without his knowledge.
  • Jeffries' office spent roughly $5,300 on AI tools from OpenAI and Anthropic since January 2025 while the leader publicly called for urgent AI regulation.
  • Pew Research found 52 percent of Americans more concerned than excited about AI, with Democratic voters registering 56 percent unease as of June.
  • AI industry groups spent $79 million on House and Senate races this cycle, the technology sector's largest election investment to date.
  • Experts say using AI while regulating it poses no inherent conflict if offices maintain transparency and proper data governance.

A staffer in House Minority Leader Hakeem Jeffries' office built an artificial intelligence program that monitored hundreds of social media accounts and generated talking points for the Democratic leader, according to details first reported by The Washington Sun.

The tool, named Caucus Pulse, analyzed more than 114,000 posts over roughly six weeks from House lawmakers and journalists. Built primarily using Anthropic's Claude AI model, the program tracked sentiment within Democratic caucus factions, flagged attacks on Jeffries, identified which members defended him, and drafted one-liners he could deploy against Republicans.

In one September memo, after a former Anthropic researcher publicly resigned with warnings about AI safety, Caucus Pulse advised Jeffries to pursue "hearings/oversight" rather than legislation. The program also suggested he avoid echoing the researcher's concerns given Anthropic's donations to Democrats, The Washington Sun reported.

Jeffries' office said the leader was unaware of the program and never used its output. Spokesperson Christie Stephenson characterized it as "a short-lived experiment with AI in a personal capacity" by a single staffer, adding that no official resources were used and the matter is being addressed internally. The office disclosed spending approximately $5,300 on OpenAI and Anthropic tools since January 2025.

Why it matters

The disclosure arrives as Democrats prepare to control both chambers of Congress starting January 3, with Jeffries positioned for a House speaker vote and the party having campaigned heavily on AI skepticism. Jeffries spent the fall calling for urgent congressional action on AI and dismissing President Trump's voluntary safety framework as "entirely unenforceable." The gap between regulatory rhetoric and internal tool adoption highlights the practical tension facing lawmakers who must both understand and oversee rapidly evolving technology.

Public sentiment shifts against AI

Voter attitudes have grown more cautious. Pew Research Center found 52 percent of U.S. adults more concerned than excited about AI in daily life as of June, up from 37 percent in 2021. Democrats registered even higher unease at 56 percent. Gallup reported 39 percent of Americans now believe AI does more harm than good, compared with 31 percent the previous year.

On Election Day, NBC News exit polling showed nearly one-quarter of voters called AI and data centers crucial to their congressional vote, with another third rating it a major factor. Those voters broke for Democrats 64 percent to 34 percent.

The use-versus-regulate debate

Experts interviewed by Newsweek saw no inherent contradiction in using AI while writing rules for it, provided offices maintain transparency.

"Members of Congress owe it to their constituents to constantly challenge themselves about how to leverage new technologies to serve them and the broader public better, while simultaneously legislating on the regulation of those industries," Nathan E. Sanders, a data scientist affiliated with Harvard's Berkman Klein Center, told the publication.

Hillary Lehr of Higher Ground Labs, which invests in progressive political technology, argued responsible use requires transparency and data privacy protections, not avoidance. She compared the situation to lawmakers driving cars while mandating seatbelts or running TV ads while passing disclosure rules.

Sanders noted that tracking public posts and sentiment with AI represents standard practice, similar to "social listening" software campaigns have deployed for years. His concern centers on data security—sensitive constituent information should only touch platforms meeting strict governance standards.

Campaign spending and industry influence

AI industry-linked super PACs invested $79 million across 68 House and Senate races this cycle, according to Tech Influence Watch data cited by Fortune. OpenAI President Greg Brockman and Andreessen Horowitz funded Leading the Future, which backed candidates favoring light regulation, while Anthropic contributed $20 million to Public First, supporting lawmakers pushing stronger safeguards.

Conservative outlets and progressive critics both seized on the Caucus Pulse story, with Truthout noting Jeffries received $150,000 from Anduril executives and $44,300 from a Palantir executive.

A spring survey by the American Association of Political Consultants found 64 percent of Republican consultants used AI daily versus 49 percent of Democrats. Republican committees embraced AI-generated ads openly, while Democrats kept the technology in back-office functions for drafts and data analysis.

Details of the Caucus Pulse program and its capabilities were first reported by The Washington Sun, which located the source code in a public GitHub repository.

  • #ai regulation
  • #hakeem jeffries
  • #congressional ai use
  • #anthropic claude
  • #political technology
  • #ai policy

This is an original analysis by the Omega editorial team. Source reporting: AI Watch.

Want systems like this working for your business?

Book a Call

More in Policy