OpenAI models being used by hostile nation states to influence operations

From Love Business East Midlands: 2024-07-07 04:00:00

A report reveals threat actors linked to hostile nation states are leveraging OpenAI tools to influence operations, with at least 5 disrupted by OpenAI. Nation states are using AI tools to advance cyber capabilities, generating text, images, articles, and social media posts. One case study, ‘Doppelganger’, exposed a Russian threat actor posting anti-Ukrainian content online using OpenAI tools. Targeting European and North American audiences, these threat actors interact with fake content to boost visibility. Despite limited positive engagement, the report highlights the dedication and capabilities of hostile nation states in influencing audiences. With the increasing development of AI tooling, threat actors are expected to enhance their use for content and code development. This poses a challenge for detection by end users and security teams. The activities discussed were related to the ongoing conflict in Ukraine, but the use of AI tools could potentially target internal issues in the UK, such as policing, to shape negative opinions.



Read more at Love Business East Midlands: OpenAI models used in nation-state influence campaigns