Israeli firms among those experimenting with AI-powered social media bots — NYT

Report on Chinese, Iranian state influence efforts says two Israeli firms used open-source models to create fake accounts; one says it was researching how to combat misinformation

By ToI Staff 20 September 2026, 11:22 am

Edit

This combination of photos shows logos of X, formerly known as Twitter, top left; Snapchat, top right; Facebook, bottom left; and TikTok, bottom right. (AP Photo)

Two private companies in Israel used open-source AI models to conduct or simulate influence campaigns on social media, The New York Times reported on Friday, in an article that also reported on use of the technology by the Iranian and Chinese governments.

One of the firms, IntelEye, said its activity – which concerned domestic Israeli politics ahead of the October election – was a research experiment, intended to bolster defenses against misinformation. Maor Sellek, one of the company’s co-founders, said IntelEye used the Chinese AI company DeepSeek for the experiment.

“IntelEye does not operate influence campaigns,” Sellek told the Times. “Our work with DeepSeek was conducted for defensive research and testing, to understand how models with insufficient safeguards could be misused and to improve detection of that activity.”

The campaign, which reportedly came to the attention of US intelligence earlier this year, saw AI agents operate about a thousand accounts on social media platforms, posting content and responding to other users.

The Times noted that some of the IntelEye bots were posting in support of Prime Minister Benjamin Netanyahu, while others were posting against him.

The other Israeli campaign mentioned in the report was referenced by Meta last month as a “commercial operation based in Israel” using a “fairly sophisticated setup,” according to the Times.

However, no further details were provided on that effort, including the firm’s name, and the Times did not appear to have any independent information about that campaign.

People talk near a Meta sign outside of the company’s headquarters in Menlo Park, California, March 7, 2023. (AP Photo/Jeff Chiu)

The article discussed the Israeli AI efforts in conjunction with state-backed influence campaigns by Iran and China.

They and the Israeli campaigns used Chinese open-source AI technology to create agents that would independently post, without significant direction by humans.

The efforts were noticed by US intelligence personnel as well as social media companies, according to the report.

The Iranian campaign – previously reported by Axios – was “especially concerning” in the eyes of officials and researchers, the Times said. Some 80,000 people followed the AI accounts, which posed as ordinary Americans in an effort to sway journalists’ and politicians’ assessments of public opinion.

In January, Channel 12 reported that about half of the politically active accounts on Israeli social media are bots. A similar situation exists in other countries around the world, according to the network.