HeadlinesBriefing favicon HeadlinesBriefing.com

AI Text Detection: Accuracy And Ethics Under Scrutiny

Financial Times Companies •
×

When Max Spero reads online content, he trusts his intuition before running it through Pangram, the AI detector he co-founded in 2023. The former Google software engineer looks for information density, arguing that AI-generated text lacks the intention behind every human word. Pangram has become central to high-profile controversies, including the Shy Girl novel removal and the accusation of Trinidadian writer Jamir Nazir at the Commonwealth Short Story Prize.

Substack chief executive Chris Best supports using the tool to scan newsletters, claiming AI content passed off as human "pollutes the commons." However, detection software faces significant criticism. Universities like Oxford, Cambridge and Harvard support AI use but do not endorse commercial detectors, noting that AI tics appear in human writing and false accusations carry severe stigma. Writers accused of using AI push back, arguing paranoia has led to witch-hunts.

Nazir maintains he did not use AI, comparing the technology to Eris, the goddess of discord. The debate mirrors historical concerns over plagiarism, but the speed and volume of AI generation make distinguishing human from machine text increasingly difficult in the modern era.