Canaries are Dying?! Quick, someone tell Stonedahl!! Warning Signs of the AI Robot Apocalypse
Let’s pretend that society is a
coal miner. On its quest to fulfill Santa’s order for 10,000,000,000 pounds of
coal (A lot of students were bad and forgot to put comments in their maze robot
code), our coal miner (society) travels deeper into the mine. Fortunately for
our coal miner, he is accompanied by his beautiful pet canary whose song fills
the air with joyous chirps in an otherwise dark, damp, and dirty mineshaft. All
of a sudden, his canary’s song turns into a cacophonous medley of daffy duck
noises and before long the bird is dead. Cause of death: super-intelligent
robots.
Much like the coal miners of the 20th
century who used to bring canaries into the mines to serve as carbon monoxide
indicators, today, people studying AI have similar fears as to what threats
will emerge as we continue to develop smarter robots and algorithms. In an article
published in the MIT Technology Review, “How to know if artificial intelligence
is about to destroy civilization”, Oren Etzioni dives into many of the
milestones that might indicate an impending Robot Apocalypse.
In the article, Etzioni refers to
AI as baby penguins, “To say that machines learn is like saying that baby penguins
know how to fish. The reality is, adult penguins swim, capture fish, digest it,
regurgitate into their beaks, and place morsels into their children’s mouths.
AI is likewise being spoon-fed by human scientists and engineers.” I thought
this quote was fantastic, as it illustrates quite clearly how much human work
is behind the development and implementation of artificial intelligence, and
how unable AI is to be able to solve problems that have not first been
formatted, digested and regurgitated into small bits by humans. This is where
his first indicator for dangerous, super-intelligent robots is introduced: the
ability of to not only take in nourishment, but also swim, hunt, chew, and
digest. As soon as robots are able to identify problems, gather raw data, and make
sense of incomplete data without any outside human involvement, they are 1 step
closer to taking over the world.
The next canary that Etzioni talks
about is self-driving cars. The argument here is that driving is such a dynamic
activity in which your surroundings are always changing and it requires a
human-like level of versatility and instinct. As of right now, robots cannot
perform this task as well as humans, but as soon as they can, it is only a
matter of time before they are able to perform a variety of other tasks
previously thought to be unique to humans.
This brings us into another theme:
the versatility of AI. For the most part, Artificial intelligence applications
are very specific and limited and rarely can be applied to or carried over to
other problems without extensive adaptations performed by human engineers. Etzioni
refers to AIs as “idiot savants,” which I think is a fair description. In order
for AI to surpass humans as the superior beings on earth, they will have to develop
versatility like humans. For example, take the maze solver we made and imagine
that instead of right angle turns, we introduce curved edges. Our robot would
have no idea what to do and would probably get quite angry at us if robots
could get angry. (Instead, to display its discontent it would probably revert to the red text bump-sensor warning
that pops up in the console that is the universal sign for “I will punch you in
the face” in robot culture). I don’t know about you, but I would be worried if
Google Maps could not only navigate roads, but also flirt with my girlfriend. Imagine
if he could not only find a path to Walgreens, but also a path into my girl’s
heart…
Although
Etzioni and myself have the opinion that AI being an extinction-level threat is
pretty far off at the moment, others feel like AI (facial recognition in
particular) is an immediate threat to humanity and freedom. The company “Clearview
AI”’s app has recently been banned by apple because of how unregulated the
field of facial recognition technology is. To learn more about how skeptical
people are about accepting AI technologies, read these articles:
This comment has been removed by the author.
ReplyDeleteI get your point but what about the Asimov's Three Laws of Robotics that we discussed in class? Is there any way to write our code, which is essentially the brain of they robots, to adhere to the above rules?
DeleteInteresting remarks about Google Maps you have there. Regarding people trust in AI, do you think it's more the manufacturer that we don't trust, because of all the security breach and data misuse?
ReplyDeleteI think our trust in the manufacturer is a significant thing. Like Johnathan said, we put trust in AI and i think our generation trusts AI more than they do human, which in the AI world are manufacturers. However, this might differ based on your perspective. But I think yeah because of security breach and data misuse are some of the reasons we find it to trust the manufacturer.
DeleteI agree with you that it's hard to see AI gain ability to change from thing that it was trained to something new. For example, it hard to see AI that was taught how to play baseball, becoming professional in basketball. The AI has it's limits and until it overcomes those limits, we are safe from them taking over the world. However, that doesn't stop other humans from using AI intelligence in wrong way. Like with facial recognition.
ReplyDeleteAs much as I agree that there is skepticism surrounding AI, I think as it grows it will be integrated into our society whether we like it or not. Thinking about historical examples, loom weaving was one of the first examples of technology replacing human workers, and although there was outrage at the time, I doubt anyone today is upset about mechanical looms. Convenience eventually trumps skepticism and I believe that years down the line the number of people who are upset about facial recognition software we turn into those against mechanical looms once the technology becomes old news and is central to everyday life.
ReplyDeleteI would have to agree with you there. Technology is just growing so fast that skepticism will not be able to outweigh technology. Already people were upset about facial recognition, but now it's almost a standard on the phones as a lock type.
Delete