DropSlop research · The House’s press releases are going AI. The Senate’s aren’t.

Key findings

Every release of each year. Before ChatGPT, the detector’s hits are false alarms; they set the yardstick each later year is measured against.

Check your own text

Paste a paragraph and see which phrases read machine-made. Free, no account. To read a whole site page by page, run a site audit.

Releases since ChatGPT from House offices and from Senate offices, each measured against its own releases from before.

A release usually quotes the member. We scored the quote and the rest of the release apart.

Members in their first two years in office, against everyone who has served longer, at the time each release went out.

Exploratory Releases of 150 words or more, by the length of the body. Each group is measured against its own releases from before ChatGPT.

Exploratory Known chatbot words, per million words of release, before ChatGPT and since. A word counts only with 20 uses before ChatGPT across 10 offices.

By the member’s home state: big delegations on their own where they reach our bar, every other state in its census region.

Releases sorted by what they announce, read from the headline and opening.

How we did it

Detector flags, with its false alarms taken out and its misses put back. The full method is on its own page.

Use these findings freely with a link back to this page.Every figure links to the data file it comes from.

Does your release read like a chatbot?

Paste a draft before it goes out, and see which phrases read machine-made.