From NBC News: 2024-05-30 18:03:52

OpenAI and Meta uncovered covert influence operations by STOIC, a firm in Tel Aviv, manipulating online political discourse. Both tech giants banned accounts and pages linked to STOIC, aiming to disrupt misinformation ahead of the U.S. presidential election. Concerns grow over AI propaganda as 20 tech companies unite to prevent election interference.

STOIC used OpenAI models to generate fake content, including fake engagement on social media. Profiles used AI-generated photos and bios, posing as different people on social media platforms. They targeted audiences in the U.S. and Canada, posting about the Israel-Hamas conflict, praising Israel’s actions and criticizing pro-Palestinian groups.

New accounts kept appearing despite platforms identifying and removing older ones. The operation struggled to engage authentically outside its network. Russia and China were also detected using similar tactics with potential AI integration. OpenAI highlighted human error in AI operations, emphasizing the importance of understanding human limitations in covert influence actions.



Read more at NBC News: Meta, OpenAI say they disrupted Israeli company’s influence campaign