Tech giants warned against AI-enabled scams targeting Consumers

For over a cen­tu­ry, the U.S. Fed­er­al Trade Com­mis­sion has tak­en action against com­pa­nies that exploit con­sumers through fraud­u­lent practices. 

The FTC has fined retail­ers that manip­u­lat­ed online reviews, made it eas­i­er to can­cel unwant­ed sub­scrip­tion ser­vices, and blocked decep­tive ads. On Mon­day, Michael Attle­son, head of the FTC’s adver­tis­ing divi­sion, warned that emerg­ing gen­er­a­tive AI tech­nolo­gies like Chat­G­PT and Dall‑E 2 could be weaponized to vio­late the unfair prac­tices pro­hib­it­ed by FTC law.

“FTC pol­i­cy holds that an act is unfair if it caus­es or is like­ly to cause sub­stan­tial injury to con­sumers that can­not rea­son­ably be avoid­ed and is not out­weighed by coun­ter­vail­ing ben­e­fits to con­sumers or com­pe­ti­tion,” Attle­son said.

New AI chat­bots like Bing, Bard and Chat­G­PT can manip­u­late users’ “beliefs, emo­tions and behav­iors,” Attle­son not­ed. These sys­tems are already being deployed as sales nego­tia­tors and ther­a­py bots with­in Wal­mart’s sup­ply chain, with the express goal of influ­enc­ing oth­ers. There is also “automa­tion bias,” where peo­ple trust AI sys­tems and see them as under­stand­ing allies. “Peo­ple may think they’re con­fid­ing in some­one who com­pre­hends them and has their back,” Attle­son said.

Attle­son acknowl­edged that the issues sur­round­ing gen­er­a­tive AI extend beyond the FTC’s man­date, but pledged that the com­mis­sion would not tol­er­ate deceit­ful com­pa­nies exploit­ing these tech­nolo­gies to take advan­tage of con­sumers. “Firms explor­ing nov­el uses of gen­er­a­tive AI, such as per­son­al­iz­ing ads for spe­cif­ic indi­vid­u­als or groups, will like­ly engage in unfair and harm­ful con­duct like finan­cial offers, in-app pur­chas­es, and attempts to unsub­scribe users,” FTC attor­neys warned.

The FTC’s rules also gov­ern how ads are placed with­in AI appli­ca­tions, unlike Google’s inser­tion of ads in search results. “Peo­ple should know if AI prod­uct respons­es are direct­ing them to web­sites, ser­vice providers or prod­ucts for com­mer­cial pur­pos­es,” Attle­son stat­ed. “And peo­ple must cer­tain­ly real­ize if they’re con­vers­ing with a human or a machine.”