The impact of using AI-powered technology for lie detection in negotiations.

• Applicants: Professor Dr.-Ing. Ayoub Al-Hamadi; Professor Dr. Joachim Weimann
• Subject Area: Economic Policy, Applied Economics
• Term: since 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 468478819

The use of AI to detect lies is an emerging topic attracting interest from research and business communities. This project aims to further develop the technology while exploring its implications. The rapid development of AI has raised many questions, particularly concerning how well science and society are prepared for this rapid development. Timely research into such technologies' consequences has become all the more important. This also applies to areas in which AI has not yet established itself, but where it is becoming apparent that it will do so - such as lie detection. Studies show that machines are already better than humans at detecting lies, especially as they are better at picking up on non-verbal cues. However, there are questions regarding humans' acceptance of such technologies. A phenomenon called "algorithm aversion" describes the reluctance of many people to accept algorithmic decisions, even when they are more accurate than human judgments. Current trends in research focus on hybrid approaches that combine human and algorithmic elements, as well as on text-based lie detection. However, there are still many open questions, especially regarding non-verbal communication and the adaptation of algorithms to different situations. Therefore, we want to answer the following questions in this project. Does algorithm aversion prevent the use of lie-detection AI? How does the use of lie-detection AI affect behavior in negotiation situations? Can the vital signs and hand gesture analysis modalities be successfully used for lie detection in video conferencing scenarios? How effective is a multimodal system that also incorporates facial expressions and acoustic modalities? At present, these questions cannot be answered profoundly. Overall, there is therefore a great need to further research these questions to better understand the impact of AI-driven lie detection on society and the economy. This proposal will help to deepen understanding in this area and investigate potential impacts.

• DFG Programme: Research Grants
• Co-Investigator: Dr. Dmitri Bershadskyy