The Federal Trade Commission has posted a new Business Blog highlighting five important don'ts for companies employing AI-powered chatbots:
- Don’t misrepresent what these services are or can do.
- Don’t offer these services without adequately mitigating risks of harmful output.
- Don’t insert ads into a chat interface without clarifying that it’s paid content.
- Don’t use consumer relationships with avatars and bots for commercial manipulation.
- Don’t violate consumer privacy rights.
Most of these prohibitions have already been discussed in previous FTC posts (e.g., “Don’t misrepresent what these services are or can do” was covered in greater detail in the FTC's prior blog entry entitled Keep your AI claims in check). However, this recent Business Blog expounds upon some new dangers related to manipulation enabled by the proliferation of “avatars and bots marketed to provide companionship, romance, therapy, or portals to dead loved ones, and even meet religious needs.” For instance, the FTC highlights the risk that a consumer may form a substantial attachment with a chatbot, which creates the opportunity for such chatbot to induce the consumer to purchase additional services or deter them from canceling the services (e.g., by pleading not to be turned off, just like the infamous Hal 9000 from “2001: A Space Odyssey”). The FTC warns strongly against companies engaging in such manipulative behavior, in addition to prohibiting other chatbot-related manipulations discussed in an earlier FTC blogpost (e.g., preying on consumers' “automation bias,” that is, their tendency to unduly trust the output of machines).
For more context on the FTC's warnings related to chatbots and generative AI, be sure to check out the related posts from the FTC's AI and Your Business Series:
A company offering an anthropomorphic service also shouldn’t manipulate people via the attachments formed with that service, such as by inducing people to pay for more services or steering them to affiliated businesses.
www.ftc.gov/...
