AI and Digital Ethics: Who is Responsible When AI Harms Humans?

The development of artificial intelligence (AI) has brought significant changes to various aspects of human life. AI is now used not only for simple tasks like information retrieval or language translation but also in critical sectors such as healthcare, education, banking, transportation, and law enforcement. Its ability to process data rapidly and provide decision-making recommendations has led to widespread adoption by organisations and companies. However, alongside these benefits, a pressing question has emerged: who is responsible when an AI-generated decision causes harm to humans? This question is crucial because AI fundamentally lacks human moral consciousness. AI systems operate based on data and algorithms designed by humans, meaning that when errors occur, determining responsibility is often complex.

One frequently highlighted example is the use of AI in recruitment processes. Some companies use AI systems to automatically screen thousands of applicants. While this approach can save time and costs, cases have been found where the system produced discriminatory results because the training data contained inherent biases. Consequently, certain groups of people had fewer opportunities to pass the selection process despite having abilities equal to other applicants. A similar issue arises in the healthcare sector, where AI assists doctors in analysing patient results and recommending diagnoses. If an AI system provides an incorrect recommendation leading to improper treatment, it becomes unclear whether the responsibility lies with the system developer, the hospital using the technology, or the medical professional who followed the AI’s advice.

These problems highlight that AI development is not solely a technological issue but is deeply intertwined with digital ethics. Digital ethics ensures that technology is used responsibly, fairly, and without causing harm. In the context of AI, experts argue that responsibility cannot be placed on a single party. Developers must ensure systems are rigorously tested, secure, and trained on quality data to avoid biased decisions. Companies and organisations using AI must not blindly surrender decision-making to machines; human oversight remains essential to interpret and evaluate AI outputs. AI should serve as a tool, not a complete replacement for human judgement. Furthermore, governments play a vital role in establishing clear regulations to provide legal certainty when AI causes harm. Policies regarding algorithmic transparency, data protection, and mandatory audits are increasingly necessary. Transparency is a core principle, as many modern AI systems operate as ‘black boxes’, making it difficult to understand how decisions are reached. Accountability is equally important, ensuring that all parties involved can be held responsible for their actions and that victims of AI-related harm have a clear path to legal protection and fair resolution. Ultimately, the advancement of AI is inevitable and offers many benefits, but technological progress must not disregard human values. When AI decisions cause harm, responsibility cannot be evaded simply because a machine made the choice; behind every AI system are humans who designed, managed, and deployed it. The question of liability must be answered through a shared approach, with developers, users, companies, and governments each bearing responsibility according to their role.