engblogs

A technical case study detailing a covert AI-driven influence operation around Ghana's 2024 election, illustrating AI-generated content, fake engagement across Facebook, X, Instagram and YouTube, and a branded website used to push partisan messaging while masking inauthentic activity.

A technical analysis of a covert influence operation that used AI-generated content and fake engagement to sway Ghana's 2024 election, revealing the Empowering Ghana network and its links to DigitSol.

A technical case study on AI-assisted deception in recruitment, detailing the North Korea–linked scheme that staged fake job applicants, used VPNs and remote tools to craft resumes, profiles, and interview responses, and leveraged social engineering across LinkedIn and hiring processes to fraudulently place individuals in Western companies.

Explores how AI-assisted translation powers a multi-channel fake-review scam ecosystem, detailing recruiter and mentor personas, the workflow from remote job offers to activation fees, and implications for detection and policy.

A technical case study of a China-origin network dubbed 'Peer Review' that deployed AI to draft surveillance-tool pitches, analyze documents, translate and review content, edit and debug code for social-media monitoring, and provide insights to authorities, illustrating the capabilities and governance challenges of AI-assisted surveillance.

A technical case study detailing OpenAI's disruption of an AI-assisted scam network that used Urdu-English translation to operate fake review jobs, exposing recruiter and mentor personas, multi-channel messaging, and an activation-fee withdrawal trap.

Case study of a China-linked AI-assisted surveillance operation, detailing how a cluster used chat models for research, code debugging, and social-media monitoring to feed intelligence and promotional materials.

An in-depth technical exploration of AI-assisted deception in the recruitment lifecycle, tracing fake resumes, support personas, social engineering, and AI-generated interview responses, while detailing TTPs and cross-industry detection efforts to disrupt North Korea-linked schemes.

Technical analysis of a North Korea-linked, AI-assisted deceptive recruitment scheme, outlining fake resumes and personas, social-engineering content, and how detection and disruption measures banned accounts and thwarted the operation.

Enhancing multi-step reasoning abilities using Twisted Sequential Monte Carlo for efficient generation of high-quality solutions.

Exploring the capabilities of the OpenAI o3-mini System Card

Exploring the groundbreaking capabilities of OpenAI's o3-mini model