August 17, 2024 at 03:03AM
OpenAI disclosed the ban on ChatGPT accounts associated with an alleged covert Iranian influence operation targeting the U.S. presidential election. The operation utilized social media and websites to disseminate content but garnered minimal engagement. Microsoft also highlighted similar threats from Iranian and Russian networks. Google’s TAG detected Iranian-backed phishing efforts aimed at high-profile users in Israel and the U.S.
After reviewing the meeting notes, the key takeaways are as follows:
– OpenAI banned a set of accounts linked to an Iranian covert influence operation named Storm-2035, which used ChatGPT to generate content focusing on the U.S. presidential election, among other topics.
– The operation utilized social media accounts and websites to share generated content, but OpenAI reported that the content did not achieve significant engagement.
– The content covered various topics including U.S. politics, global events, conflicts in Gaza, Israel’s presence at the Olympic Games, and the U.S. presidential election.
– Microsoft also highlighted Storm-2035 as an Iranian network engaging U.S. voter groups with polarizing messaging and warned of an increase in foreign malign influence activity targeting the U.S. election from Iranian and Russian networks.
– The group set up phony news and commentary sites, utilizing AI-enabled services to plagiarize content from U.S. publications.
– In response to enforcement efforts, the propaganda network has shifted tactics, using non-political posts and spoofing entertainment news outlets to evade detection.
– Meta reported uncovering six new influence networks from Russia, Vietnam, and the U.S. in the second quarter of 2024 and noted a decrease in Doppelganger’s attempts at sharing links to its domains.
– Google’s TAG detected and disrupted Iranian-backed spear-phishing efforts aimed at compromising the personal accounts of high-profile users in Israel and the U.S., attributed to a threat actor known as APT42 and its state-sponsored hacking crew affiliated with Iran’s Islamic Revolutionary Guard Corps (IRGC).
– APT42’s phishing attacks utilized social engineering techniques and abused services such as Google, Dropbox, and others to gather credentials from Google, Hotmail, and Yahoo users.
Overall, the meeting notes outlined the activities of an Iranian covert influence operation using AI-generated content, as well as the responses and warnings from organizations like OpenAI, Microsoft, Meta, and Google regarding foreign malign influence activities and spear-phishing efforts.