# Digital Twins as Funhouse Mirrors: Five Key Distortions

arXiv: https://arxiv.org/abs/2509.19088 · HTML: https://mlsanity.com/paper/2509.19088/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Sep 2025.

## Built on by (most cited first, 11 of 11)

- [TextBO: Bayesian Optimization in Language Space for Eval-Efficient Self-Improving AI](https://mlsanity.com/paper/2511.12063/) (2025)
- [SYN-DIGITS: A Synthetic Control Framework for Calibrated Digital Twin Simulation](https://mlsanity.com/paper/2604.07513/) (2026)
- [Large Behavior Model: A Promptable Digital Twin of the Retail Customer](https://mlsanity.com/paper/2607.06993/) (2026)
- [When Can You Trust Your Synthetic Users? Diagnostics and Corrections for LLM Consumer Panels](https://mlsanity.com/paper/2609.13148/) (2026)
- [The Digital Twin Counterfactual Framework: A Validation Architecture for Simulated Potential Outcomes](https://mlsanity.com/paper/2604.01325/) (2026)
- [Rectification Difficulty and Optimal Sample Allocation in LLM-Augmented Surveys](https://mlsanity.com/paper/2604.17267/) (2026)
- [Adaptive Querying with AI Persona Priors](https://mlsanity.com/paper/2605.00696/) (2026)
- [Generating Public Health Responses using Survey-Augmented Large Language Models](https://mlsanity.com/paper/2606.21820/) (2026)
- [Modeling Human Behavior with Type Vectors Using AI](https://mlsanity.com/paper/2608.18265/) (2026)
- [Seeing Is Not Perceiving: When Synthetic Consumers Can and Cannot Pretest Visual Marketing](https://mlsanity.com/paper/2609.25677/) (2026)
- [AI-Moderated Interviews for Market Research and Digital Twins Calibration](https://mlsanity.com/paper/2609.29143/) (2026)

## Abstract (arXiv)

Scientists and practitioners are increasingly moving to deploy digital twins--LLM-based models of real individuals--across social science and policy research. We conduct 19 pre-registered studies spanning 164 diverse outcomes (e.g., attitudes toward hiring algorithms, intentions to share misinformation), comparing human responses to those of their corresponding digital twins, which are trained on each individual's prior responses to over 500 questions. We establish an empirical benchmark for digital twin performance: their predictions are only modestly more accurate than those of a homogeneous base LLM and exhibit weak correlation with human responses (average r = 0.20). To inform future development, we identify five systematic distortions in digital twin behavior: (i) insufficient individuation, (ii) stereotyping, (iii) representation bias, (iv) ideological bias, and (v) hyper-rationality. Finally, we release our full dataset and code as a standardized testbed for evaluating and improving digital twin methodologies. Together, our findings caution against premature deployment while laying the groundwork for a transparent, replicable, and iterative science of responsible digital twin development.