Use of AI Tools for Forensic Purposes: Ethical and Legal Considerations from an EU Perspective

Artificial Intelligence (AI) may be a useful tool for enhancing the work of forensic scientists, but it also raises crucial ethical and legal concerns, which are the focus of this chapter. Our view is that, firstly, AI should never be considered a substitute for “human-based forensics”. This is because it is often necessary for a court to ask questions around the nature and the methods undertaken to carry out forensic examinations and when the subject matter is technically complex. This is especially true when we consider the lack of transparency of these tools, which can make it challenging to understand the reasoning behind the conclusions they lead to reach. This is a significant obstacle in a judicial debate. Therefore, it is necessary to include a human element in the process, a forensic expert capable of answering any questions that may arise from the use of these AI tools. If this requirement cannot be met, the use of AI would be illegitimate, as it would violate the defendant's right to defense. Secondly, the fact that AI tools will inevitably contain biases is cause of concern. The decisive factor is whether or not these biases are admissible. A bias that belongs to an unimportant category differs from a bias that belongs to a relevant category. An AI bias that is similar to that which is common in human-performed forensic science may be permissible, unless sensible arguments are put forward as to why an increased demand should be placed on the machines. Indeed, in modern times, the comparison to the human standard should indicate the acceptable range of bias. On the other hand, it is important to remember that AI can also be used to reduce bias in traditional forensic science by acknowledging, to begin with, that human decisions are influenced by cognitive bias.