Abstract
The authors discuss Large Language Models in the context of the security risks associated with their functions and availability. Even though their applications seem to be similar to search engines and internet access, the true danger posed by Large Language Models lies in basic analytical and programming skills they provide to any criminal or terrorist. They assert that accessible Large Language Models not only diminish financial barriers to various criminal activities but also lower the expertise and commitment required by individuals or small groups to commit crimes, and acts of terror in particular. On the other hand, however, law enforcement agencies can also harness the capabilities of these models to stay ahead of emerging threats.
Publisher
Uniwersytet Jagiellonski - Wydawnictwo Uniwersytetu Jagiellonskiego
Reference37 articles.
1. 1. Breakstone J. et al., Students' Civic Online Reasoning: A National Portrait, "Educational Researcher" 2021, no. 50, pp. 505-515. https://doi.org/10.3102/0013189X211017495.
2. 2. Europol, ChatGPT. The impact of Large Language Models on Law Enforcement, Luxembourg 2023.
3. 3. Faesen L. et al., Red Lines & Baselines Towards a European Multistakeholder Approach to Counter Disinformation, The Hague Centre for Strategic Studies 2021.
4. 4. Felson M., Cohen L., Human ecology and crime: A routine activity approach, "Human Ecology" 1980, no. 8, pp. 389-406. https://doi.org/10.1007/BF01561001.
5. 5. Fuocco M.A., Trial and error: They had larceny in their hearts but little in their heads, "Pittsburgh Post-Gazette" 1996.