Artificial intelligence

Abstract: Artificial Intelligence is becoming increasingly embedded in various areas of human life, offering new capabilities that go beyond traditional software systems. Unlike conventional programs that follow fixed instructions, AI can generate its own solutions after processing large volumes of data. However, human input remains essential in designing AI architecture and setting its goals. While AI improves efficiency and decision-making across fields, it also introduces new types of risks. These risks often arise not from malicious intent, but from unpredictable system behavior and user errors. This paper analyzes such risks using a systems perspective and logistic S-curve modeling to examine the AI lifecycle. The analysis shows that the first three stages—development, scaling, and stabilization—carry the highest levels of vulnerability. Key issues include design flaws, insufficient debugging, and lack of continuous monitoring. More advanced systems may evolve through multiple S-curve phases, each introducing new challenges. The study emphasizes the need for stronger legal and ethical standards, drawing on regulatory efforts from the EU, USA, UK, Germany, and France. International cooperation is also highlighted as a key factor in ensuring that AI develops safely and responsibly.

Keywords: artificial intelligence, Al-system, risk, vulnerability, system lifecycle, S-curve

References:

1. Crawford K. (2021) Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. Yale University Press. 336 p.
2. Russell S.J., Norvig P. (2003) Artificial Intelligence: A Modern Approach. 2nd ed. Prentice Hall.
3. Luxton D.D. (2014) Artificial intelligence in psychological practice: Current and future applications and implications. Professional Psychology: Research and Practice. V. 45, I. 5, P. 332–339. https://doi.org/10.1037/a0034559
4. Yudkowsky E. (2006) Artificial Intelligence as a Positive and Negative Factor in Global Risk. Draft manuscript. [Online]. Available: https://intelligence.org/files/AIPosNegFactor.pdf
5. Hutson M. (2018) Missing data hinder replication of artificial intelligence studies. Science, February 15. https://doi.org/10.1126/science.aat3298
6. Bejan A., Lorente S. (2011) The constructal law origin of the logistics S curve. Journal of Applied Physics. V. 110, 024901. https://doi.org/10.1063/1.3606555
7. Reed T.R., Reed N.E., Fritzson P. (2004) Heart sound analysis for symptom detection and computer-aided diagnosis. Simulation Modelling Practice and Theory. V. 12, I. 2, P. 129–146. https://doi.org/10.1016/j.simpat.2003.11.005
8. The Medical Futurist. (2016) Artificial Intelligence Will Redesign Healthcare. [Online]. Available: https://medicalfuturist.com/artificial-intel-ligence-will-redesign-healthcare
9. Yorita A., Kubota N. (2011) Cognitive Development in Partner Robots for Information Support to Elderly People. IEEE Transactions on Autonomous Mental Development. V. 3, I. 1, P. 64–73. https://doi.org/10.1109/TAMD.2011.2105868
10. Fawcett T. (2004) ROC Graphs: Notes and Practical Considerations for Researchers. Kluwer Academic Publishers.
11. Davis J., Goadrich M. (2006) The Relationship Between Precision-Recall and ROC Curves. Proceedings of the 23rd International Conference on Machine Learning, Pittsburgh, PA. P. 233–240. https://doi.org/10.1145/1143844.1143874
12. Liang D., Tsai C.-F., Wu H.-T. (2015) The Effect of Feature Selection on Financial Distress Prediction. Knowledge-Based Systems. V. 73, P. 289–297. https://doi.org/10.1016/j.knosys.2014.10.011
13. Luger G., Stubblefield W. (2004) Artificial Intelligence: Structures and Strategies for Complex Problem Solving. 5th ed. The Benjamin/Cummings Publishing Company. 720 p.