As the use of artificial intelligence within industry and society as a whole continues to grow, so does the call for it to reflect our ethical values and ensure that when it is used, it is done so with consideration of these.
Ethical AI can broadly be split into two areas – The Ethics pertaining to artificial intelligence and artificial intelligence with ethics. The latter is often referred to as machine ethics, with the former, known as roboethics. The latter is not to be understood, however, as exclusively for robots, since ethical issues are prevalent within all areas of artificial intelligence application.
The ethics of technology is a longstanding research area, however, new issues arise when considering artificial intelligence. Traditionally computer programs were made, and the computers were explicitly told what actions to carry out, by simply programming requests. However, with the introduction of approaches such as machine learning computers are being allowed to make decisions on their own, which brings a new range of ethical problems. Such ethical questions include:
- Which data should I collect and how should it be managed?
- What are the implications of my AI system?
- Who will the system impact, and what is their understanding of the technology being used?
- How can we mitigate against decision errors?