People will accept AI systems only if their outcome is seen as meaningful with a low risk level.
– AI researcher and entrepreneur Josef Baker‑Brunnbauer, Social Tech Lab.eu
Trust is both a moral property and a feeling that is crucial for our ability to operate as communities and societies. Public power cannot act in a way that undermines people’s trust in the administration or one another.
One challenge of new technology is the increased amount of chatbots and automated services, even in the public sector. Trust as an experience is created through respectful interaction where everyone is basically on an equal footing.
Can we develop machines that can make people feel seen, heard and appreciated?
Or would it be wiser to provide AI services that do not even attempt to simulate a feeling of mutual trust between people? These kids of services would deliberately retain their machine-like features so that people would not mistakenly project any human features on them.
Current systems represent narrow artificial intelligence. This refers to a situation where an individual AI focuses on solving or optimising a specific task assigned to it. The impacts of such AI systems are often also examined in a narrow fashion, from a perspective that only focuses on the task of the AI.
If you involve more of the groups that could be unintentionally affected by your system already at the design stage, you will have an easier time anticipating different impacts in the development. This way, you could reinforce positive impacts and prevent negative impacts.