Skip to content
context Regulation and law

US Department of Justice deploys detection tool Pangram against AI-generated legal filings

only one source so far

The DOJ is using the detection tool Pangram to uncover legal filings generated with ChatGPT. In a voter fraud dispute, government lawyers accused the defense of "defense by ChatGPT" and of filings containing nonexistent precedents.

The US Department of Justice (DOJ) is using the detection service Pangram to identify text it believes was generated by artificial intelligence in filings submitted by the opposing party. According to court documents reviewed by 404 Media, this came to light in a high-profile voter fraud case, where government lawyers accused the defense of running the case as "defense by ChatGPT" and of submitting filings containing "AI slop" – meaning low-quality text generated by artificial intelligence.

According to 404 Media, the government lawyers wrote in their filing that litigating the case felt, from their perspective, like responding to documents generated by a robot attempting to practice law, and that the defense used what "children would call AI slop" to produce repeated filings that, in their view, needlessly burden the court. The disputed filings are also said to include hallucinated legal cases that do not actually exist.

404 Media has long documented cases in which judges exposed lawyers using AI after they cited nonexistent precedents, and in which the lawyers were even financially sanctioned for it. The new development is that both sides of a dispute, including government lawyers, are now resorting to detection and mutual accusations of AI use within a single case.

What changed

Why it matters

The case shows that AI text detection is becoming part of the court battle itself – parties to a dispute are now throwing it at each other as an argument. For the legal profession and for companies using AI to prepare documents, this means a real risk that insufficiently verified content will be exposed and used against them, including sanctions for citing nonexistent cases.

Two audiences, two different impacts

What this means

01

For individuals

Lawyers and people working with legal documents should expect that using AI without verifying the content can be detected and used against them in court.

What to do When writing or reviewing legal filings, verify whether the cited cases and arguments actually exist, and do not rely on unchecked output from AI tools.
More practical updates →
02

For a business

For companies and legal departments, there is a growing risk that insufficiently verified use of AI in drafting filings will be exposed by detection tools used by the opposing party or the court, which can lead to sanctions and reputational damage.

Risks and compliance
What to decide Review internal procedures for using AI in preparing legal and court documents, and introduce human verification of citations and cases.
More business impacts →
ChatGPT AI detection DOJ hallucinations Pangram legal documents

Check the original

Event sources

only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.

1
404 Media (investigative tech journalism) independent context · first detected 'Defense by ChatGPT'