The question is not whether AI can replace us but what role we play in ensuring that AI does not irreversibly harm humanity.
The use cases of AI are continually evolving. We have seen a recent explosion in large language model (LLM) capabilities that have barely scratched the surface. But alongside an exploration of how AI can help us in our daily lives, significant investments are being made across new areas, such as AI for military capabilities. The future of responsible AI, however, cannot exist without humans.
The time spent thinking we may become obsolete provides the perfect opportunity for humans to realise how vital we are to the future; to exercise all the things that make us human such as reflexivity and reflectivity, critical thinking and emotional cognition. We must understand that the harms and biases proliferated by AI are a mirror of the existing fractures that exist already.
AI is a tool and if extra efforts are not taken to amend the existing biases which it may have picked up through design, training data or lack of robustness, it could result in harm at unprecedented speeds.
In The Netherlands, for example, 1.4m people were wrongfully flagged as fraudulent by an AI-powered risk profile system due to bias, which resulted in tens of thousands of people being pushed into poverty and thousands of children mistakenly taken into foster care.
The question then becomes not whether AI is going to replace us but what role we as humans play in ensuring that AI does not promote harm or have irreversible effects on humanity.
