OpenAI says it has banned a cluster of ChatGPT accounts tied to a coordinated criminal scam operation that used the chatbot to draft fraudulent messages, create social media comments and translate scam material into multiple languages. The enforcement action matters because it shows how organized fraud networks can use consumer AI tools to expand familiar schemes across platforms and borders.

According to OpenAI, its investigators linked the accounts through a review of activity associated with the alleged scam network. The company said the operators used ChatGPT at several points in their workflow, including writing unsolicited pitches, producing comments intended to build credibility and adapting material for targets who spoke different languages.

The reported activity followed a well-established fraud pattern. Targets received unexpected messages offering easy earnings or investment returns, after which operators allegedly tried to move conversations to another service and direct victims toward a fraudulent earning or investment platform. Requests for fees or cryptocurrency could then be used to extract money, OpenAI said.

OpenAI reported that its models refused some requests submitted by the operators. The company also said the underlying behavior remained recognizable despite the use of more fluent, personalized and multilingual text. After identifying the activity, OpenAI disabled the associated accounts and shared its findings with industry partners that could search for linked users or conduct on-platform reviews.

AI Expanded a Familiar Scam

The case suggests that generative AI can help fraud networks produce convincing messages more quickly, personalize pitches and reuse scripts in multiple languages. Those capabilities can reduce the cost of reaching potential victims and weaken a common warning sign: awkward or inconsistent writing in unsolicited messages.

OpenAI’s account also indicates that AI did not create a new form of fraud. The network allegedly applied the technology to an existing social-engineering sequence: unsolicited contact, promises of easy money, relocation to another channel and referral to a controlled platform. That recurring pattern may give technology companies an opportunity to detect coordinated behavior earlier, before a bank, cryptocurrency service or police agency receives a complaint.

Inside the Network’s Workflow

The operation allegedly began with AI-written outreach offering earnings or investment opportunities. Operators then used generated social media comments to create the appearance of activity or approval around the scheme. OpenAI said translation tools allowed the same fraudulent playbook to target several countries without relying on native speakers for each market.

The disclosure forms part of OpenAI’s broader effort to police scams and other malicious uses of its models. Removing accounts can deny operators access to one service and provide useful indicators to other platforms. It does not establish that the wider criminal organization, its communication channels or its financial infrastructure has been dismantled.

Operators and Victims Remain Unnamed

The public account comes from OpenAI’s own investigation, and the source material provides no independent confirmation of the network’s identity or reach. OpenAI has not named the alleged operators, disclosed their location or said how many accounts were removed.

It is also unclear how many people received the messages, whether anyone lost money or how much revenue the operation generated. No arrests or criminal charges were announced in the supplied material, and OpenAI did not say whether law-enforcement agencies are involved. The identities of the industry partners that received information and the status of any related accounts on their platforms also remain undisclosed.

Partner Reviews May Extend Enforcement

Industry partners can now review the indicators shared by OpenAI and determine whether linked accounts or infrastructure appeared on their services. Further disclosures could show whether the enforcement spread beyond ChatGPT or led to the removal of fraudulent profiles, messaging accounts or payment channels.

OpenAI may also publish more details through later threat reports, including the network’s scale, techniques or geographic focus. Until then, the confirmed public action is limited to ChatGPT account bans and information sharing; the broader scam operation’s current status remains unknown.

Key Questions

What action did OpenAI take?

OpenAI says it identified and banned a cluster of ChatGPT accounts associated with the coordinated scam operation. It also shared findings with industry partners.

How was ChatGPT allegedly used?

The operators allegedly used it to draft fraudulent outreach, create comments intended to build trust and translate scam scripts into several languages.

Did the AI create a new type of scam?

No new fraud method was described. OpenAI said the network used AI to expand a familiar sequence involving unsolicited contact, promises of easy returns and attempts to direct targets to fraudulent platforms.

Were any operators arrested?

No arrests were announced in the supplied source material. It is not clear whether law enforcement is investigating the operation.

Has the criminal network been dismantled?

That has not been established. The bans removed access to one technology platform, but OpenAI did not say whether the operators’ wider accounts, payment systems or fraudulent sites remain active.

Source: Thorsten Meyer AI