Overview
- Billionaire Stanley Druckenmiller’s Wall Street Journal op-ed was flagged as 100% AI by the detector Pangram and he publicly acknowledged using AI to draft the piece, and Journal opinion editor Paul Gigot said the paper will not police contributors’ AI use.
- A Semafor analysis using Pangram found a limited but real presence of AI in guest essays at the New York Times, Wall Street Journal, and Washington Post, labeling 10 of 310 recent submissions as at least 80% AI and another 40 as partially AI-generated.
- Detection tools like Pangram report low false-positive rates but researchers warn they can misclassify work by non-native English writers and produce errors, which complicates using these tools as final proof of AI authorship.
- Major outlets hold different rules—The New York Times bars AI drafting of guest essays, The Washington Post requires authors to confirm no AI was used, and the Financial Times restricts AI in writing—but enforcement has been uneven, especially for high-profile contributors.
- The episode has sharpened two central stakes: whether publications will require clear disclosure to preserve reader trust and whether market demand will normalize AI-assisted opinion pieces, with critics also warning of risks such as reduced originality in arguments.