OpenAI has reported an attempted phishing attack by a group called SweetSpecter, suspected to have ties to China.

This attack aimed to compromise the personal data of OpenAI employees, raising new concerns over cybersecurity threats facing leading AI companies amid an increasingly tense US-China relationship.

In early 2024, SweetSpecter impersonated ChatGPT users, sending deceptive customer support emails to select employees.

These emails contained malware attachments that, if opened, could enable the hackers to capture sensitive information and exfiltrate additional data.

OpenAI’s security protocols prevented these malicious emails from reaching employee inboxes, and the team informed targeted employees about the spear phishing attempt.

This incident underscores ongoing cybersecurity risks for US tech firms as China’s state-linked groups exploit espionage tactics to gain sensitive insights into artificial intelligence technologies.

Earlier this year, a former Google engineer was charged with stealing AI trade secrets for a Chinese entity.