Brian X. Chen, The Times's lead consumer technology writer, spent a week testing Pangram, an AI text detector that costs $20 a month. The tool calculates a percentage score indicating how much of a pasted block of text was generated by a chatbot versus written by a person. In dozens of experiments, the detector never failed at distinguishing chatbot output from human writing.
Pangram is one of many AI detectors that have hit the market in recent years, but it stands out for its accuracy. The company says its technology can correctly identify AI-generated text 9,999 out of 10,000 times. An independent study by the University of Chicago also found the tool was near-perfect with text, a sharp contrast to earlier scanners that made glaring errors, including misidentifying writing by Charles Dickens as machine-made.
The scale of the slop problem
Chen tested Pangram by feeding it personal essays alongside text generated with chatbots instructed to imitate his writing. He also uploaded social media posts from LinkedIn and Reddit that he suspected were AI-generated. The detector passed every test.
The timing matters. Some studies say 50 percent of online articles are now artificially generated. At that rate, AI detectors could become a staple app, akin to antivirus software, that people use to shield themselves from phony, low-quality information online.
Chen described the motivation behind the testing plainly: "Recently, as I was scrolling through my phone, I realized I was suffering from a new kind of fatigue: slop-induced burnout. Almost all the posts I was reading on social media were riddled with bullet points, emojis and em dashes. They clearly looked like the work of a chatbot."
Where detection still falls short
Pangram's weakness is images. Like other AI detectors, it isn't very good at identifying AI-generated pictures, which Chen noted are a more potent medium for spreading misinformation online.
That limitation matters for anyone whose work involves verifying content. Text detection is improving quickly, but the same reliability doesn't yet extend to visual media.
Why this matters for writers
For writers, tools like Pangram offer a practical way to check whether content you're reviewing - or content submitted to you - was actually written by a person. That's useful for editors vetting freelance pitches, hiring managers reviewing test assignments, or journalists assessing sources. But the image gap is a reminder that detection tools are not a complete solution. Writers who rely on these tools should treat them as one layer of verification, not a final judgment. The cost may also be a factor: at $20 a month, Pangram is priced for professionals who regularly need to verify authorship, not casual users. For those working in fields where AI-generated content is a growing concern, training in how these tools work and where they fail is becoming part of the job. Resources like AI for Writers and AI for Technical Writers offer practical grounding in the technology behind these detectors and its limits.
Your membership also unlocks:
