Pangram, an AI text detection company, has faced backlash for its scores being misused to shame social media users.
The company's CEO, Max Spero, continues to use Pangram scores to call out alleged AI use on social media, despite earlier cutting ties with a journalist who helped publicize the scores.
This practice has led to accusations that the scores imply a lack of personal effort or originality, which is not what the tool measures.
Pangram's tool only detects whether AI was involved in content creation, not how it was used. However, users have turned these scores into public shaming, suggesting that any AI involvement equates to dishonesty or laziness. This misinterpretation has led to situations where individuals who used AI for editing or language smoothing are unfairly criticized.
According to WIRED, journalist Rod Breslau was hired by Pangram to publicly shame users for AI use. Breslau justified his approach by stating that many professionals on social media claim to be leaders but do not write their own posts. He argued that using AI for content creation should be met with public scrutiny.
Pangram's business model relies on the stigma of AI use as laziness or dishonesty. If this stigma fades, there may be less demand for AI detection services. The company now aims to detect even light AI editing, positioning itself as an authority on originality checks in publishing, education, and beyond.
The company's approach has led to situations where academic papers are rejected based on AI scores, even when the AI was used to improve clarity for non-native speakers. This highlights a gap between detection and the actual intent behind AI use.
There is a historic analogy to Alexandre Dumas, whose work was aided by assistants, yet the final product reflected his vision.
Pangram has not commented on the specific implications of its scores beyond detection, and the company has not addressed the broader issue of how users interpret its results. The company continues to use its scores to hold individuals accountable for alleged AI involvement, despite the limitations of its detection method.
Source: thedecoder
