HEADER
Amazon Tested AI Designed To Recruit
New Employees And It Went Horribly Wrong
SUMMARY
The new starts explaining that the
first AI was programmed and set loose in social media. As a result it ended up
turning into Nazism and white supremacy. As a failure it was taken offline.
Since that failure, Amazon had been
working on another AI programmed to recruit new employees. It was designed to
choose the best applicants out of a significant amount. Lots of people worked
on this project until they finally had it. At first it seemed a success, the AI
started learning how to deal and organize information about applicants and choose
the best options. The truth was that it was prioritizing men over women. It
learned the words men used most commonly and prioritized them over women’s.
Also it downgraded candidates coming from all-women colleges and penalized
resumes containing the word “women’s”. The AI was eventually shut down, it was
yet another failure.
An Amazon spokesperson said it was
never used to evaluate candidates, but in a trial and development phase.
Instead, a weaker version helps the agency, while a new recruiting AI has been
commissioned, aimed at increasing diversity.
Technology specialists say that more
work needs to be done on machine learning.
LINK OR REFERENCE
PUBLICATION DATE
11 October 2018
SCIENTIFIC SCOPE
Technology
TEXT COMMENTARY
Even though technology is developing
at a really fast speed, the AI experiments and its results show us that the
human race isn’t capable yet of creating or programming something similar to a
human life artificially using technology.
I think that the main problem here
lies in the capacity of reasoning and thinking, as Turing test can show. We
human can learn, which is similar to the program AIs have, but reasoning is not
something you can learn, it comes naturally. So the machines learn human
thinking and reasoning to think and reason themselves. Since the human race is
corrupted, the machines end up being corrupted from the information they receive.
That’s why experts think more work
needs to be done technologically, since something is still leaking, though they
probably mean it not just at a technological level, but also at a social level.
GLOSSARY
AI: Artificial Intelligence
No hay comentarios:
Publicar un comentario