Meta reveals 'likely AI-generated' deceptive comments praising Israel's handling of the Gaza conflict
Meta has identified 'likely AI-generated' deceptive content on Facebook and Instagram, which includes comments that praise Israel's approach to the Gaza conflict. These misleading comments were spotted under posts from various global news outlets and statements from U.S. lawmakers.
Deceptive Accounts and Campaign Details
According to Meta's quarterly security report, the deceptive accounts masqueraded as Jewish students, African Americans, and other concerned individuals, primarily targeting users in the United States and Canada. The campaign has been linked to STOIC, a political marketing firm based in Tel Aviv, Israel. STOIC positions itself as a comprehensive solution for managing campaigns of all sizes.
As per the company's LinkedIn profile, STOIC offers services such as data enrichment, online and offline micro-targeting, digital asset management, and more.
Utilization of Generative AI
This marks Meta's first acknowledgment of the use of text-based generative AI in influencing operations. The social media giant had previously detected AI-generated profile pictures in similar scenarios starting from 2019.
Generative AI technology can rapidly and inexpensively produce content that closely resembles human-created text, images, and audio. However, researchers have expressed concerns about its potential to facilitate the spread of misinformation and impact electoral processes.
Response and Assurance from Meta
Meta's security executives confirmed that they swiftly dismantled the Israeli campaign and emphasized that the introduction of new AI technologies did not impede their efforts to disrupt such influence networks.
Despite the speed and scale at which generative AI tools can generate content, Meta's detection capabilities and countermeasures remained unaffected. Mike Dvilyanski, the Head of Threat Investigations at Meta, stated, 'There are several instances within these networks where generative AI tools are likely used to produce content. While it may enable them to create content more rapidly or in larger quantities, it has not compromised our ability to identify them.'
This article contains information sourced from Reuters.
