The AI text detection tool from the company Pangram is used to publicly ridicule users
The company Pangram hired a journalist who used its detection score to publicly ridicule users suspected of using AI. It ended the collaboration, but CEO Max Spero continues to publish scores – yet the tool measures only AI involvement, not the extent of human contribution.
The company Pangram operates an AI text detection tool, and its founder Max Spero hired journalist Rod Breslau this year, who was to use the tool’s score to publicly ridicule social media users suspected of using AI. Breslau said he wanted a “tougher approach” because he felt that people using AI had “gotten off too lightly”. The company Pangram later ended its collaboration with him and, according to CEO Spero, now expects AI use to become increasingly accepted and wants to position itself as an authority on checking originality in publishing, education and other areas. Spero himself, however, continues to publicly share scores and point out alleged AI use by specific people.
According to the source, the main problem is that the score from the Pangram tool measures only whether AI was involved in the text in some way, not how – whether it involved generating the entire text from a prompt, using AI as a thinking partner, getting help with wording, or simply polishing or translating the author’s own text. According to the author, public ridicule based on the score implies that the person did not think about the text themselves or work on it – something the score alone does not demonstrate. According to the text, the business model of the company Pangram depends on people associating AI use with laziness or dishonesty.
As an example, the author cites his own English version of the article, which was translated from a German original created with the help of AI and subsequently edited both manually and with AI – the Pangram tool rated it as “28 percent AI” and flagged the final paragraphs as text written by AI, even though they had been produced using the same process as the rest of the text. According to the author, academic institutions are already using similar scores to reject scientific papers, including those by researchers who write in a foreign language or are better at science than at writing and use AI to express their findings more clearly.
Why it matters
The detection score does not distinguish between generating an entire text with AI and using it as an assistive tool (refining wording, translation), so public accusations based on it may also affect people who came up with the text themselves and only edited it with the help of AI. According to the source, this already extends to academic institutions that use the score to reject papers, which may disadvantage authors writing in a foreign language.
Two audiences, two different impacts
What this means
For individuals
The detection score does not distinguish between text generated by AI from start to finish and text that the author wrote themselves and only edited or translated with the help of AI – public ridicule based on such a score may therefore also affect honest work.
For a business
According to the source, academic institutions are already rejecting papers based on the percentage score from the Pangram tool, which, according to the author of the text, also affects authors writing in a foreign language or authors who used AI only to refine wording, not to create content.
Risks and complianceCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.