Simulated Artificial Intelligence Is Imminent

Robots are advancing at a extraordinary rate.  They are doing so in many dimensions.  Boston Dynamics' Atlas is more agile than most humans.  

 

Now U.K. based company, Engineered Arts is setting new standards for realistic facial expressions.                   

 

Obviously, manual dexterity is also central to a successfully functioning general purpose robot.  Most of these research efforts have been directed at prosthetics but they can also be applied to robots.

                            

People tend to focus on the idea that a robot might become 'conscious' and, thus, be more similar to a human.  However, that is really not the most functional product feature for human use.  That, after all, is why, if they are, robots will be built and bought.  

The core of the artificial intelligence component will look like a juiced up Alexa.  If you ask your robot to play a game of chess with you, through the onboard natural language program it will access an Internet chess program and play the game.  If you ask for Coq au vin it will find a recipe on the Internet and prepare it for you.   If you ask your robot to get you a beer out of the frig, it probably doesn't need the Internet.  However, by piggy backing off of the knowledge easily found on the Internet, your robot will appear to be nearly omniscient.

The Kurzweil crowd is fascinated with 'AGI' The 'Singularity' is a time when they imagine that a superintelligence, functionally incomprehensible to humans will emerge.  It is generally taken to likely happen between 2040 and 2050.  It is odd that in this 'Singularity' community to which many high profile people have belonged, the notion of more or less unending exponential growth in technology is credulously accepted.  This is odd because a review of the history of technology strongly suggests that technologies progress sigmoidally or on an S-curve.  For example, if one plotted the maximum speed attainable by humans as an exponential curve, we should be exceeding the speed of light by now.  Of course, we are not even close to that.  In the 20th Century, the growth in maximum speed hit a point of inflection and has been slowing down ever since.

It is important to recognize that it is extremely difficult to estimate the inflection point or the asymptote in most cases.  We have good reason to believe that the 'Moore's Law' https://www.synopsys.com/glossary/what-is-moores-law.html exponential growth has reached its point of inflection.  However, the development of simulated intelligence in computers and robots is more complex than just Moore's Law.  There are other technological, economic and software components that affect the emergence of Artificial General Intelligence.  Still, a point at which humans become secondary to computer intelligence is not a foregone conclusion.  

There is also the matter of what may exist in the lab and what is implemented in the general society.  Robots and computer based AI, as well, are first and foremost a product.  There are, for many reasons, a huge inventory of technologies, demonstrated in the lab, that do not get commercialized because they are not attractive products.  It is very likely that a humanoid personal assistant robot, as an advancement upon Alexa and its competitors, would be a very attractive product, even if it was very expensive.  However, there is no clear benefit to making them 'conscious' and there are reasons to not give them any form of independent volition.  They are a product as a servant and servants are intended to have very limited initiative.  

In the above example, if I tell my robot to get me a beer out of the frig, I don't expect it to start lecturing me on the calories and refuse to do so.  If I asked my robot to play chess with me and I instruct it to play at a FIDE of 1,000, I would not expect it to up its rating when it started to lose.  If it did such things, I would take it into the shop and ask that it be 'fixed'.  A robot that is conscious, reflective and self-motivated is like a flying car.  Even if it is technically feasible, it won't be commercialized because it is a bad product.  

I do not know if the super-intelligent, conscious AGI is even possible, but, if so, it is not likely to exist outside of labs for quite some time.  However, we are right on the verge of creating a humanoid robot that is an extremely capable personal assistant.  A rudimentary version is probably feasible right now simply by creating a joint venture between Boston Dynamics, and Engineered Arts and an inclusion of a few subcontractors.  The first commercially available robot will likely be introduced prior to 2030.  It will likely have limited sales at a 300K 2022USD price point.  However, like most emergent technologies, over the following decade, the price will likely fall significantly and, perhaps to the point where they become common in middle class households.

When I was young, fantastical technologies were limited to Science Fiction.  Because of Arthur C. Clarke, Isaac Asimov, et alia, as I reached young adulthood in the late 1960s, more practical and rigorous analysis of future technologies began surfacing.  However, there was still a gee-whiz flavor to them.  Then authors, such as Alvin Toffler, started writing books purporting to present hard-boiled looks at future technologies and that tradition has remained right through to today with Ray Kurzweil and Peter Diamandis.While they report on actual research, however, they suffer from what I call, 'Everything that can go right, will go right'.  Even a preliminary review of their history demonstrates that most of what could go right, won't.

I spent my career in business and I worked in various aspects of future planning.  We were truly hard-boiled, because when things didn't turn out well, we'd get called on the carpet.  Advanced humanoid robots powered by Internet based, simulated artificial intelligence is a 'today' technology.  We could form a joint venture between Boston Dynamics, Engineered Arts and Amazon right now and start the product development process.  No basic research is needed; the technologies already exist.

Will this fundamentally change the world?  Probably.  Humanoid robot personal assistants and home domestic robots will likely be second behind only the house as the largest personal expenditures.  Of course, our homes will likely be highly automated with robotics that is not humanoid.  The robotic vacuum that wanders about our apartment today is just the first of many examples.  Essentially, household chores will become a thing of the past.