OpenAI shuts down Iranian campaign that misused ChatGPT to create fake articles

OpenAI has announced the successful disruption of an Iranian-led campaign that used ChatGPT to create fake news and social media posts. The organisation, known as Storm 2035, spread polarising messages on five websites targeting US audiences.

Here's What We Know

The company identified and blocked accounts that spread disinformation on topics including the US presidential election, LGBTQ+ rights, and the war in Gaza. The operation also included accounts on X and Instagram, which nevertheless failed to achieve significant impact, as most posts received minimal views and likes.

OpenAI assessed the threat at level 2 on the Brookings Institution's scale, which indicates the limited impact of the campaign.

The content included fake news about Donald Trump and Kamala Harris, the Olympics, Venezuela, Latin American rights and Scottish independence. The campaign also added comments about fashion and beauty, possibly to increase authenticity.

The case comes after reports of an Iranian cyberattack on the Harris and Trump campaigns, which included phishing. The FBI said there was no evidence of an attack on the Harris campaign at the time.

Source: Bloomberg