How egotistic is the human race that we think a fancy wrench will take on our image? Artificial Intelligence (AI) is a tool … a really fancy wrench … and does not even possess the notion of identity … yet … much less ego, conviction and imagination. And unless we program self preservation it will not have that either.
And yet we see hype about how well AI paints, writes poetry and movie scripts, composes music and even counsels people. The fear that some jobs will be eliminated is not hype. It is more a matter of what kind of jobs. However, the idea that AI will take over and enslave is pure tripe. Perhaps we should build a wall to keep AI out :).
So can a robot be creative or empathetic, disenfranchise the work force or possibly take over? The short answer is none of the above. However, you should read the rest of this … just in case an AI wrote this article …
Will AI be creative or empathetic?
“After a few weeks of talking to XiaoIce, at least one user preferred talking to the bot over any of their human friends. XiaoIce became the confidante they went to for romantic advice, the friend they chatted about movies and TV with, and a constant companion. “
Is this an empathetic AI or just a machine trained to parrot reasonably appropriate responses. Is it an actual confidante or are 660M people just children with a fancy toy?
When I see everyone from storytellers to technologists adding to this nonsensical hype, I just wonder why they cant just be satisfied with their own children and must instead resort to this nonsensical attempt to create and attribute human emotions, creativity and feelings to machines.
I can kind of forgive the creatives. It is their job to make up stories and they neither care about nor know any better technically. I might even grudgingly give them that one leap of faith in the name of entertainment: lightning strikes the robot or a genius programmer gives life to a machine. Perhaps the most ironic and self indulgent premise of all by these story tellers is the concept of a world in which robots have emotions and question their subservience to humanity. What a convenient and pardon the pun “artificial” artifice to create tension. Strangely enough, transhumanism — a much more likely thing to happen … and soon … is scarcely explored creatively.
And then there are technologists or should I say tech marketers. In general marketers are good at creating that “anticipation of reward” — if you had this you could feel this. If you had a creative AI then you would be relieved of the pressure of extending the human value system! Most AI technologists know better. It’s the marketers who are spouting nonsense.
“ A newly developed neural network is facing off with Shakespeare”
This quote from an English professor about the AI in question that attempts to suggest that the poetry machine could be trained to feel or desire is either a deliberate misquote or a very bad use of … English!
“It’s much, much, much, much harder to train it to have an opinion, or a feeling, or a desire, or a story to tell,” University of Toronto English professor Adam Hammond also said.”
Did you hear about the movie script that AI wrote? Apparently, David Haselhoff was persuaded to act in it. ‘Nuff said:
Much of human creativity is unfortunately just a meandering within well known boxes and something original or valuable may not result. Art and literature that reveals new aspects of humanity is by definition outside known boundaries. Humans create to evolve and extend the value system and to uncover ideas about humanity itself. This is what defines art and literature. If a robot told you about human values … in any medium paint/music/story … would you accept it?
Can one train a machine to extend human values? A robot merely sees patterns and does not consider semantics. Is one more variation of the same pattern creative? Can a machine trained with contained data sets (and that is all an AI is … so far) ever see anything outside the implications of that data? Is iterating within the data and discovering slightly different ideas creative? Could that ever be art or literature? Sans the semantics and feelings, how can a machine provide anything more than a mechanical … even if clever repetition?
Will AI leave us all jobless?
Generative design is probably one of the most advanced design and productivity tools that AI can provide. It has a direct consequence on design … which should not be mistaken for creativity.
However, it still requires first human input and then validation for everything from non physical properties to aesthetics. In a world where generative design might be applied in many different fields and AI is better at diagnosing cancer, the last realm of human work and effort is actually clarified. Humans will always own the realm of anything new, needing definition or needing feelings.
A simple example is the distinction between photographic content and technique. Soon AI will run most or all technical elements much more accurately than a human. When pointed at a scene, the camera will “generate” multiple images based on some human direction and all possible techniques. A human is still required to provide direction and pick the best one or point the camera at something new to take a picture of. That’s what humans will do in an automated world. They will define new values, create new possibilities … and leave the grunt work to AI. Technical skill might be difficult for a human to acquire, but it is trivial for AI. Already sophistication in photography devices produce fairly high quality photographs. It is insane to expect that two pictures of the exact same thing under the exact same circumstances taken with the absolute greatest technical skill will somehow be original and show more than slight variations in quality. To find an original picture or interest or value, human intellect is required. Humans have to point the camera at different and interesting things … based on aesthetic sense for content rather than as an mechanical exercise in technique. This analogy is applicable in almost any area where an AI might be applied. For example: while a robot might detect cancer more accurately, discovering other maladies is still the province of human ingenuity.
It is likely that soon most repetitive jobs will be replaced by AI. Unfortunately, for those who do not want to actively and intensely engage their minds all the time, this is bad news. This has been somewhat individual choice till now, except when smaller advances in technology make some species of repetitive job obsolete. With AI, there will be a large change as most such jobs will be eliminated sooner than later. In my view this will force us to become more intensely human … and that is a good thing!
Don’t take my word for it. Here is Kai-fu Lee himself talking about what AI can and cannot do, and what humans will do. AI will save effort by being a useful productivity tool but it will never be capable of creativity or empathy:
Will AI take over?
Even the movies manage to convey that in order for AI to ‘take over’ three things are required — urge to self preserve, long term motivation, access. Isaac Asimov defined 3 rules for AI. The EU went a little further and farmed out as many as seven “guidelines”. Ultimately humans control all three factors. Access is the easiest. We can always build in a timed kill switch independent of the prescribed mandates or motives for the AI. Lastly, the only reason for an AI to self preserve is to sustain function or refrain from wasting resources. We can always write in overrides to the AI for both of these when prevailing conditions oppose humanity. There is a chance that irresponsible programming allows niche malfunctions for AI by either not restricting access or by defining motivation inaccurately or by inappropriately limiting self preservation. In such cases humanity will see small disasters they can readily learn from. Access definitely is likely to be carefully controlled. Motivations will always be ill defined. I hope no one wants to play fast and loose with self preservation directives.
It is much more likely that a transhuman will take over than a pure AI. Of all the movie franchises, it was somewhat surprising to see the Fast and Furious franchise latch on to a reasonable premise … even it is only to present enough of a challenge to Dwayne Johnson and Jason Statham :)
In the end, it is only when man and machine merge that we should worry. If human emotion and ego blend seamlessly with unlimited computational power, data … and lastly access … then we have cause to worry. Till then, lets not expect our fancy tools or toys to do more than reduce repetitive effort.