In the past, this was of little importance, but now that data processing has taken on its most literal meaning, it has become extraordinarily significant.
And why is that?
Because in the “prehistoric age of information technology,” the user had full control over the data processing process, as they managed the IT solution they used through its buttons and controls. However, with the advent of AI assistants, the user loses full control over the data processing process. Their role now shifts to focusing on submitting the amount of data they wish to process, accompanied by natural language instructions, to an AI-powered program and awaiting the results.
In the past, the user was solely responsible for the outcome, as the IT program was merely a tool. Now, however, shared responsibility is emerging between the user and the AI system. Previously, an average user never had to consider the actual process of using an IT program, but today, it has become a matter of great significance. While this may seem strange for now, consider the example of a self-driving car. The driver remains responsible for the journey, but the decisions made by the car also influence the events.
AI providers largely exclude their responsibility in the general terms and conditions they apply, leaving the user solely responsible for the outcome, even though they are no longer entirely the creators of that outcome.
Since we cannot fully see into the data processing carried out by AI, and it remains uncheckable, any legal liability questions regarding the result are determined by how cautious the user was and how much they practiced good habits when initiating communication with the AI.
Where does this matter?
- Copyright disputes, where it can be difficult to determine to what extent AI-generated content is original and how much it builds upon the work of others. The user must act cautiously to avoid infringing on copyright laws and be aware that the legal status of AI-generated content is not yet fully clarified.
- Data breaches, where personal data processed by AI may end up in unauthorized hands or be misused. The user may be held responsible for the security of the data handled by the AI and must ensure that the AI complies with data protection regulations.
- Suspected discrimination or bias, where AI decisions may adversely affect certain groups. The user must monitor the AI-generated results and report if there is a suspicion of discrimination or bias.
- Spreading false or misleading information, where AI-generated content may be inaccurate or misleading. The user must critically evaluate the information generated by AI and verify its authenticity before passing it on.
- Security incidents, where AI-driven systems malfunction, potentially causing financial loss or personal injury. The user must be aware of the limitations and risks of AI-driven systems and ensure their safe use.
AI is not just a tool but a partner in solving tasks. The user must be aware of the data on which the AI makes its decisions and the ethical or security issues its use may raise. Critically evaluating AI-generated content and responsible use are essential for harnessing the benefits of the technology and minimizing potential risks.
In conclusion, it is important to emphasize that data flow is not only the technical foundation of AI operation but also the foundation of the legal and ethical relationship between the user and AI.
Dr. Kozsla László, AI Officer
Dr. Miklós Péter, GDPR lawyer
