Making Machines Moral

Machines don't have
a moral built in. We have to give that to them. In particular, when
we teach AI model based on data, – if the data that we use
for teaching, has biases, – the artificial intelligence
will also continue those biases. Those biases will be present
in their decisions. This is why we
absolutely have to build – trust and ethics into the core
of these mechanisms. Ethics is at the core
of artificial intelligence, – because it is a tool, much like
any other tool used by humans. It can be used for good and bad.
That's why we as a society – need to set the rules and
what is acceptable and what is not. Artificial intelligence itself
will not do that for us. Additionally, it is really important
to do the AI research – and implement the systems
so that they are reliable. That is also an ethical thing. Thinking for instance
of self-driving cars. It's really important they
are constructed based on – the best practices and we can trust
that they function as designed. As an example of the need
for regulation, – there are undemocratic states
out there – that employ mass surveillance
of their citizens – using extremely sophisticated
computer vision and other AI tools. This is an existing example
that shows – that we really have to set
the limits openly and in our law.

Leave a Reply

Your email address will not be published. Required fields are marked *