Better Knowledge for a Hopeful Future | Interview with William Ammerman (Part 3)


– [Narrator] In our
last video, technologist and author, William Ammerman,
spoke about the dangers of the modern era of digital marketing and how protecting ourselves requires a deeper knowledge
of new technologies. In our third and final
video, Ammerman walks us through some of the terms
we’ve heard at many times in the media, but might
not totally understand. He then gives us a hopeful story of how these technologies can
be used for the public good. – An algorithm is a set of instructions, let’s be very super clear about that. It’s a set of computer instructions. You might follow an algorithm
in your morning ritual of turning the alarm off
and getting up out of bed. Now that is an algorithm of sorts. You reach over, the alarm
goes off, you reach over, you turn off the alarm, you
lay there for a few seconds, you swing your feet out of
bed, you sit on the edge of your bed for a second, you stand up. Every step in that process could be recorded as a
step-by-step procedure. Another way to think about algorithms is that they work together,
they work reciprocally. They interact with one another. Great example. There was an auction at Christie’s where they auctioned off
this piece of artwork, and I think it sold for
something like $40,000. And it was a painting,
painted by a GAN algorithm, and this algorithm really
took the combination of two algorithms working in concert, actually kind of competitively. One algorithm was designed
to create works of art, billions of combinations that
it thought were humanlike. The other algorithm judged the output of the first algorithm. So the first algorithm is just designed to make pretty pictures. The second algorithm is designed to judge which ones humans will like,
based on machine learning. So now you’ve got two algorithms working hand in hand to produce a piece of artwork that actually sold for $40,000
at Christie’s auction house. – [Alexander] But how can these algorithms be used to better serve the public good? Ammerman offers a hopeful story. – Something very interesting happened when Wikipedia was created. Wikipedia was created
largely by human contributors who were contributing
information into Wikipedia. And, over time, Wikipedia got
populated, and somebody looked at it and said, “You know,
it’s a very strange thing, “but Wikipedia seems to be biased “in favor of male scientists.” They did a study on
this, and they realized that, yeah, women, female scientists are vastly underrepresented in Wikipedia. And they started to go backwards and try to deconstruct
that, and they figured out that the majority of Wikipedia
contributors were men. And I don’t know exactly
what the percentages were, but more men were contributing
to Wikipedia than women. And so, they actually
wrote an algorithm to go in and identify this bias, and
start to correct for it. So they started to try to
intentionally balance the scales, and try to say, okay there’s lots of brilliant female scientists. They just happen to be underrepresented by a bias in the system. We’re going to actually use an algorithm to go back and try to correct that bias. So, not only can algorithms
produce biases unintentionally, but the good news is, if we
identify those algorithms, we can actually use algorithms to correct biases and identify them. – [Alexander] Thanks for
watching this installment of our deep dive into the world
of artificial intelligence. To hear more from William Ammerman on the modern age of digital marketing, be sure to check out his
book, “The Invisible Brand”. If the AI system on your smart phone can’t help you locate a copy,
then check out on Amazon, Barnes & Noble, or
anywhere books are sold. This is Alexander with UpTech Reports.

Leave a Reply

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