If we can’t teach machines to internalise human values and make decisions based on them, we must accept – and ensure – that AI is of limited use to us.

https://www.newstatesman.com/science-tech/2021/02/how-prevent-ai-taking-over-world

The so-called “Value Alignment Problem” – how to get AI to respect and conform to human values – is arguably the most important, if vexing, problem faced by AI developers today.

Stuart Russell, a leading AI scientist at Berkeley, offers an intriguing solution. Let’s design AI so that its goals are unclear. We then allow it to fill in the gaps by observing human behaviour. By learning its values from humans, the AI’s goals will be our goals.