Content authentication¶
Content authentication is the process of verifying that digital content (text, images, video, user profiles) is genuine and comes from a claimed source. Authentication operates at multiple levels: user identity (is this profile real?), content provenance (did this user write this?), and claim veracity (is what is claimed true?).
On platforms with weak identity verification (e.g., LinkedIn with only email, Twitter with only phone optional), imposters can create convincing fake accounts. As LLMs improve, the problem intensifies: detecting automatically generated profile sections becomes harder and requires sophisticated textual analysis. Authentication thus spans both profile-level (is this a real person?) and content-level (is this text human-written?) verification.
Key papers¶
- The Looming Threat of Fake and LLM-generated LinkedIn Profiles: Challenges and Opportunities for Detection and Prevention — Detects fake LinkedIn profiles using textual features (SSTE method); discriminates legitimate, human-fake, and LLM-generated profiles.
- Can LLM-Generated Misinformation Be Detected? — Shows LLM-generated misinformation is harder to detect than human-written falsehoods, raising challenges for content-level authentication.
- Survey of Hallucination in Natural Language Generation — Surveys hallucination in NLG; relevant to detecting LLM-generated content that falsely claims facts.
Related topics¶
- Fake accounts — detection and prevention of fraudulent profiles
- Fake news detection methods — computational methods for detecting fake content
- Generated text detection — distinguishing human-written from machine-generated text
- Large Language Models — generation of synthetic content via LLMs