OpenAI is investigating reports of a potential data breach affecting millions of user accounts, but says it has yet to find evidence supporting the claims.
The alleged breach, circulating on a hacking forum Friday, suggests a threat actor has compromised the login details of 20 million OpenAI accounts, including email addresses and passwords.
The individual behind the claim offered an apparent sample of the stolen data and is attempting to sell the complete dataset.
While the veracity of these claims remains unconfirmed, OpenAI has stated it is taking the reports "seriously" and is actively investigating the situation.
The credibility of the claims has not been verified, but in a statement OpenAI said it was looking into the reports.
“We take these claims seriously,” an OpenAI spokesperson said.
“We have not seen any evidence that this is connected to a compromise of OpenAI systems to date.”
The AI firm is the maker of ChatGPT, the AI chatbot which has exploded in popularity since its original launch in late 2022.
Cybersecurity expert Jamie Akhtar, chief executive and co-founder of CyberSmart, said consumers should exercise additional caution by updating passwords and log-in credentials, and warned that cybercriminals could look to exploit the incident.
“If verified, this breach could have huge ramifications, both for OpenAI and its customers,” he said.
“Millions of people and businesses have embraced the company’s technology into their daily lives, so the potential damage to OpenAI’s reputation for data security could be huge.
“Worse still, compromised accounts could be used to access and abuse sensitive customer data or to exploit OpenAI’s APIs and distribute malware and other cyber nasties.
“There’s also the possibility of cybercriminals using stolen user credentials to produce targeted phishing campaigns, steal identities, or commit financial fraud.
“Although this breach is yet to be verified by OpenAI, anyone using the tool should update their passwords and credentials, as a precaution.
“And, if you haven’t already, switch on multi-factor authentication within OpenAI’s settings, as this should give you another layer of protection even if your password has been compromised.”
