AI developers' responsibilities

AI companies have responsibilities to develop safe AI tools.

When Microsoft launched one of the first AI chatbots, Tay, in 2016 it was manipulated by users into posting racist and sexist content within 16 hours. It was shut down and Microsoft and other AI companies worked much harder to make future chatbots safer.

This is why AI companies build filters and guardrails to block harmful requests and prevent their tools from generating dangerous and inappropriate responses.

AI companies also have a responsibility to protect users' privacy.

They have to abide by the privacy laws of the country where the tools are being used. These are stricter in places like Europe and Australia than the US, for example.

They must tell users what data they are collecting and how long they are keeping that data.

Children's data gets extra protection under laws around the world.