Co-Founder and Co-Director ethix: Lab for Innovation Ethics and PostDoc researcher at the University of Zurich and EPFL answers.
When we speak about ethical AI, we usually tend to think about the use and application of AI systems. So think about maybe the biggest question in this debate is the impact of AI on the job market, on the kind of competence which is required to find a job in, let's say, 20 years, which kind of job would disappear in 10-15 years?
This is the application of AI to meet the impact of society, which we philosophers tend to consider as a social justice question.
So it's really a question of justice. Think also about "AI for good movement", which is how to develop new AI applications for good purposes. Think about health, climate change, any kind of very big challenges we are currently facing, and how can AI contribute to addressing and solving these issues?
But I would like to do one step back and think about the design of AI application AI system, which is one of the core issues we're currently discussing in the scientific discussion on AI ethics. Think about data ethics, because data is a key part of AI applications as we know them. For now. We need to think about the collection of data.
Which kind of data can we use? Which one are we allowed to use by law on one side, but also based on ethical reflection?
Think about personal data, for instance, as a case of data, which is not very clear to which extent and in which context we are allowed to use. But think also about the design of AI system and application, the way the interface between the user of the application and the programmer
of the application is conceived and made.
This is an important ethical question because it's about manipulation. It's about transparency in giving some information about how the system is working. Of course, it's about securing that the system is explainable that the user can understand the type of decision taken by the system.
Maybe last point. The very important point, if you think about integrating AI systems and applications in your team, in your company, is that an AI is not only a technical system, it's a system working with human beings, with team members.
And you need to think about the integration of the system within the team. It's about power relations, it's about decision-making processes, it's about expertise in the team.
So the quick answer to the question is, can AI be ethical? We need to consider all these different elements of the ethics of AI theory to judge and evaluate a specific case in a specific context.
If a specific AI system and application can be justified on ethical grounds.
All AI Insights Videos here.