Workado ordered to tone down claims for its 'AI detector'

Wondering whether that recipe, language course or news story you’re reading is generated by artificial intelligence (AI)? Well, you could just flip a coin. Or you could pay for an “AI Content Detector” marketed by Workado LLC.

The Federal Trade Commission (FTC)says the coin toss will be about as accurate as the supposed detection software. It wants Workado to stop making unsupported claims about theproduct.

Workado marketed its AI tool as a way for consumers to determine whether online content was written by a human or generated by AI technologies like ChatGPT. The company claimed the tool was “98% accurate,” but the FTC found that in reality, the detector was accurate only 53% of the timebarely better than guessing.

Consumers trusted Workados AI Content Detector to help them decipher whether AI was behind a piece of writing, but the product did no better than a coin toss, said Chris Mufarrige, Director of the FTCs Bureau of Consumer Protection.

False advertising and the consequences

According to the FTC, Workados AI tool was primarily trained to classify academic content, not the broader range of materials it claimed to handle, like blog posts and Wikipedia articles.

The proposed FTC order includes several key provisions:

• Workado must only advertise effectiveness claims if they are backed by competent and reliable evidence.

• The company must retain supporting evidence for any performance claims.

• Workado is required to email affected customers about the settlement.

• The company must submit compliance reports to the FTC for the next three years.

The settlement will be open for public comment before it becomes final. It reflects a broader FTC effort to hold AI-related companies accountable for truthful marketing as the industry rapidly expands.