FTC sanctions: AI

The FTC has announced proposed settlements with Cox Media Group, MindSift and 1010 Digital Works over allegations that the companies deceived customers about an AI-powered marketing service branded as “Active Listening.”

Link

According to the FTC, the companies claimed that the service could use conversations captured from consumers’ smart devices to target localized advertising, and also claimed that consumers had opted in to this type of targeting. The FTC alleged that both claims were false. The service did not use voice data from consumer conversations, and consumers had not provided opt-in consent for such targeting.

Photo: freepik Magnific

The proposed orders require the companies to pay a total of $930,000: Cox Media Group is to pay $880,000, while MindSift and 1010 Digital Works are each to pay $25,000. The money is intended to provide redress to impacted CMG customers.

For me, this case is important for several reasons.

➡️ It shows that this is a a compliance and TPRM risk.

If a company says that its product uses AI, advanced algorithms, smart targeting, behavioral intelligence or real-time data analysis, those claims must be accurate, documented and explainable. AI cannot be used as a “magic label”.

➡️ This case shows that privacy claims and consent claims must be treated as control statements. Saying that users “opted in” is not just marketing language. It is a factual representation about the legal basis and governance of data processing.

The FTC made a very clear point: clicking through mandatory app terms of service does not automatically constitute opt-in consent.

➡️ The case is a reminder that deception risk is not limited to consumer-facing notices.

The FTC alleged that the companies deceived potential business customers by overstating what the service could do. In addition, MindSift and 1010 Digital Works were charged with providing CMG with the “means and instrumentalities” to deceive customers through marketing materials, sales pitches and responses to customer questions.

During the recent Infosecurity Europe conference in London, I saw AI-related banners at almost every booth. Many products and services were presented as “AI-powered,” “AI-enabled” or “supported by AI.”

That is exactly why, during the purchasing process, organizations should ask a very practical question:

☑️ How much of this is marketing — and how much of this is actually true?

For every product or service, buyers should perform adequate TPRM due diligence.

Autor: Sebastian Burgemejster