Tip:
Highlight text to annotate it
X
The artificial intelligence arms race is on.
Who will be the first to achieve full self awareness?
Earlier this week, Mark Zuckerberg
revealed on an earnings call that Facebook
has formed an artificial intelligence group, which
will develop new AI tools using the vast collection of images
and other data that users have uploaded to the site.
According to the project's lead researcher,
the first goal will be to implement image recognition.
Meaning when you upload a photo, Facebook
will know, without you tagging it,
what it is, what's going on, or even who's in it.
That might sound like an unexciting use of AI.
But even that requires a huge amount
of computational power relying on layers of trial and error
and analyzing an image.
Google already has an opt-in "find my face"
feature for Google+, though I haven't seen it used that much
yet.
And on Google Glass, which would seem
like a natural medium for face recognition,
that tech is currently banned.
And as a side note, some designers
are already coming up with creative ways
to thwart facial recognition software.
But beyond just image scanning, the real dream of AI
is to build computers that work just
like the brain with the ability to self learn
and to teach itself new tasks.
Google, for example, is partnering with NASA
on a quantum AI project with the goal of vastly speeding up
that process.
With that quantum twist, computers
could analyze visual or other cues exponentially more
quickly.
But already existing AI tech is being
used in training simulations for military and commercial use
where users can program simulated human agents
to respond autonomously to cues.
And of course, personal assistants, like Siri,
can respond to our questions, even if they don't always
get them right, at least not right now.
And for more on current advances in robotics,
check out Trace's video on Valkyrie,
one of the latest beasts from NASA.
Thanks for watching.
See you next time.