Late last month, the Federal Trade Center (“FTC”) announced that it reached a settlement with a company called Workado to resolve allegations that it made false or misleading representations about the efficacy of its “AI Content Detector” product in violation of Section 5 of the FTC Act.
According to the FTC’s complaint, Workado advertised its AI Content Detector as trained on a number of mediums, including blog posts, Wikipedia entries, and AI-generated text, when it fact it was trained only on academic abstracts and content generated by ChatGPT and was not adequately fine-tuned or tested. The complaint alleges that, based on independent testing, the AI Content Detector is less accurate than Workado advertised and is accurate only around half the time when evaluating non-academic AI-generated content.
The proposed settlement order places a number of restrictions and obligations on Workado. including:
- Prohibiting the company from making representations about the efficacy of AI Content Detector unless such representations are evidence-based;
- Creating and retaining records of any evidence that supports representations about the efficacy of AI Content Detector and any documents relating to its testing protocol;
- Notifying all customers who subscribed to Workado’s AI Content Detector that Workado entered the settlement agreement; and
- Submitting annual compliance reports to the FTC for three years.
The proposed consent order is subject to a 30-day public comment period; any comments must be received by June 2, 2025.