OpenAI recently released a report outlining disturbing abuses of ChatGPT. These include romance scams, influence campaigns, and cybercrime.
The company decided to publish the findings for security reasons after uncovering alarming misuse of its technology.
Romance Scams and Dating Frauds
The report highlights how scammers exploited ChatGPT to carry out dating frauds, particularly targeting men in Indonesia. Over a hundred people fell victim to these fake online romances.
Scammers used ChatGPT to create convincing ads and messages, making fake content for dating sites. The goal was to trick victims into completing tasks that promised easy money.
In some cases, scammers even posed as lawyers, pressuring victims to send money. They relied on AI-generated chats to make their schemes seem more legitimate.
Influence Campaigns and Political Targeting
The report also reveals that some of the banned accounts tried to run influence campaigns. One targeted Japan’s first female prime minister.
Although OpenAI did not provide details on how these campaigns operated, it confirmed that these efforts were calculated and coordinated.
Fake profiles posed as policy experts or paid consultants, emailing US state officials to influence decisions.
Data Collection and Cybercrime Concerns
OpenAI’s report also points to instances of cybercrime. Banned accounts used ChatGPT and other tools to gather sensitive information.
They targeted data on Americans, online forums, and the locations of federal buildings. Some accounts even attempted to learn how facial recognition software works raising serious security concerns.
OpenAI’s Response to the Misuse
OpenAI has acted swiftly in response to these threats. All accounts involved in the misuse have been banned.
As AI grows more powerful, OpenAI acknowledges that the risks increase. The company is working closely with authorities and strengthening safety measures to prevent further misuse.
