Session 6: AI Models are not perfectionists. What's the right framework to decide who should be responsible for their mistakes? Is it the Banality of Evil?
The past quarter has seen lots of changes to the global world order. The United States has formally exited the World Health Organization, AI models have drastically increased in complexity since the start of the year and adoption of AI has also increased. The hype around AI has raised the question regarding who is to blame if AI Models make mistakes.
With regards to this SAYDF organized a discussion on the idea if Arendt’s Banality of Evil concept is relevant to the field of Artificial Intelligence (AI). At the start of the session most participants made it clear whether they agree or disagree with Arendt’s philosophical concept. Participants in favor of the concept gave examples of complicated AI models and they have recently been in the news for wrong reasons. They highlighted how it is tough to pinpoint who is to blame - the creator of the chatbot or the user - for these atrocities as a result of this complex nature. As a result of this they blame the user. Furthermore, they also stated that people have recently started to trust AI more due to a rise in its accuracy. They argued that this cannot be used as an excuse to the repercussions of the use of AI on the AI Model. This is as humans actively decided to trust the result of the model. At the same time, other participants believed Arendt’s concept is not relevant to the field of AI. This is as they argued that the user of the AI was not following any command and they had just asked the AI Model for assistance. Furthermore, they also said that in times of pressure when someone who is desperate asks for help they are not thinking critically, which further causes the Banality of Evil concept to not apply to them.
The discussion challenged participants. It made them grapple with questions they’d not initially thought about as the Banality of Evil is usually associated with more traditional atrocities such as conflicts. However, despite this both sides presented arguments which showed their critical thinking skills.
Links for further interest
Comments