Machine Behavior: Season I – Trailer

Machine Behavior: Season I – Trailer


That’s why the machine-behavior
program is really needed, not just to bring in
nonscientists and engineers, but also to help enhance the
tool box of the scientists and engineers. We’re evolving robots, and this
was the craziest and coolest idea I’d heard of. Something people are calling
autonomous, moralistic agents. So you mean there are
self-driving cars everywhere. Because big data is not just
about humans, of course. The technology and the
human have to work together. Right now our communication
occurs online. Algorithmic trading
in financial markets. Be amplifying the biases. How is the car going to
make ethical decisions? Usually when we
think of censorship, we think about fear. What are the mechanisms that
will bring about cooperation? How does it need
to engage a person to empower that person to kind
of become who they want to be, right– not doing it for
them but supporting them so that they can do
it for themselves. I don’t really like to
think about the viewpoint that it’s machines with
lots of information and lots of computational capabilities
kind of competing with people. I always like to really
think more about machines representing people. I continue to be interested
in developments in AI and how it’s going
to change the world, and I kept hearing about
the autonomous cars making morally
charged decisions. How do you make sure that
the current data you have is not reflecting human
biases in them in ways that, when a model
learns from data, it’s not going to be
amplifying the biases we have in our society? I think there are
ways that algorithms can be used against
censorship, and then there are ways that
algorithms can be used for the purposes of censorship. What if everything is– the
internet is online, right? Then it’s a completely
different calculation because if that is the case,
even your offline decisions are actually really mediated
by what the online world wants. Of course we’re being
affected by the way that our machines learn,
and there’s definitely a feedback process by which
we design our machines to do something and
then the consequences that we now learn
differently because of that. And let’s look at
the dating example. So it’s possible that
with the availability of these apps
which you can carry with you and this
extraordinary variety and number of individuals
that you come to believe are possible candidates
for dates with you. There are a lot of interesting
research questions. Also, this question,
what is the role? I don’t know that AI
has an answer to it. In the long run, what would
be good for the species and good for us? I prefer to be optimistic. I’m certainly concerned. I’m kind of a realist
and a bit of a pessimist. It’s a very interesting
question to think about what’s going to happen there. These are questions that are
very difficult to answer. Whether or not this is good
for democracy, I’m not sure. [LAUGHS] That seems like a hell
of a thing to do, right? All of these things
are augmenting people and the communication
between people. This is the moment that we
need to be working on trying to ensure the better future.

Leave a Reply

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