Job Description
Rigorous social-science reasoning depends on careful use of evidence, theory, and methodology — the same discipline AI systems need to learn before they can be trusted with sociological, political, or anthropological analysis. Chegg is seeking social science professionals to evaluate AI-generated analyses, build authoritative reference answers, and flag flawed reasoning in AI-produced content. Fully remote and asynchronous.
Core Responsibilities
- Evaluate AI-generated responses to sociology, political science, and anthropology questions for accuracy, reasoning quality, and completeness
- Write reference-standard answers with correct theoretical framing and evidence use with correct established social-science theory and research methodology and clear explanations that AI models learn from
- Identify factual errors, misapplied theory, or unsupported claims in AI-produced content, tagging the specific error category
- Compare and rank multiple AI responses to the same problem, explaining which best reflects sound methodology and balanced interpretation
- Test whether AI systems handle controversial or politically sensitive topics while maintaining neutrality and evidentiary rigor
- Create realistic synthetic case studies and research scenarios across social-science subfields used to train and evaluate models
- Contribute to rubric design — defining scoring criteria such as accuracy, theoretical grounding, and clarity — used to benchmark AI performance at scale
- Complete asynchronous task batches independently — no scheduled calls or fixed hours required
Key Qualifications
- Master’s degree or PhD in Sociology, Political Science, Anthropology, or a related social-science field (or equivalent recognized qualification); relevant experience is a plus — postgraduate and doctoral graduates welcome
- Strong grounding in social-science research methodology, theory, and academic writing conventions
- Ability to evaluate arguments for evidentiary support and logical consistency
- Sharp eye for subtle factual, logical, or methodological error — can detect and explain flawed reasoning in plain English
- Self-directed, detail-oriented, and comfortable delivering quality work asynchronously
Nice to Have
- Published academic research or teaching experience at the university level
- Proficiency with Stata or R for causal inference methods (difference-in-differences, instrumental variables, regression discontinuity)
- Experience with GIS for spatial analysis or Qualtrics for survey design
- Prior experience in content review, QA, or AI/ML data annotation projects
Why Chegg
- Fully remote and flexible — no fixed schedule
- Task-based commitment, typically 10–40 hours per week
- Real-world challenges with direct impact on AI accuracy and safety
- Pathway to ongoing projects for high-quality contributors