Recently much press has been devoted to the fact that the worlds leading GO game champion was defeated by a computer. These single discipline achievements are not uncommon. As far back as the 80s MIT developed a computer program called Mathematica that solved differential equations better than ANY human being. Since then numerous of these more headline grabbing AI consequences have come out periodically. This, in my opinion is no more impressive than saying a 10 ton robot can lift 100 times more weight than the strongest human. Some will say that is an unfair comparison but I think not.
Recent Advances Deep Learning
The most impressive thing that AI has done in the last few years is demonstrated by a technology called “Deep Learning” which is a version neural networks developed back in the 60s originally. Recent advances have made these deep learning nets able to outperform any other type of AI significantly allowing computers to demonstrate almost human levels in some new areas such as vision, speech recognition and even answering questions on Jeopardy. This has created a lot of press lately and excitement. I have talked about how we are still far from AGI (General human level intelligence) and ASI (super intelligence greater than humans) in this blog here.
Many people are assuming that until computers get to AGI (General human level intelligence) or ASI(Super human intelligence) computers and AI won’t be a problem. We are very far from AGI or ASI. However, the statement that we don’t have to worry about the computer or AI till then is NOT true. I have tried to point out that like the mechanical machines have had a major impact on human productivity and labor the ANI type AI we have developed can have a similar massive changes in labor without needing to be human-level intelligence or better.
The crux of the matter is how many of our jobs really involve creativity?
The main jobs at threat from this technology are jobs that don’t involve creativity or don’t involve having to switch from one specialty of knowledge to another or combine specialties of knowledge. Jobs that don’t require humor for instance which is one of the most creative things people do. We have never successfully produced programs which create humor or understand humor. Surprisingly there has been significant effort put into teaching a computer to recognize satire. This has been found to be an extremely hard problem. The purpose of this would be for instance in reading people’s emails or decoding their speech to detect if they said “blow you up” means really blow you up or is meant in some other context that is benign.
Many jobs have a significant amount of rote activity in them. Some jobs are purely rote. In the past as our technology improvements have come along they have destroyed some jobs for physical labor. People no longer break stones to make bricks for homes (except for maybe some high end homes.)
Physical labor involving significant strength is not needed but even more complex consturction activities such as building a brick wall is automated to an astonishing extent already today.
He have giant construction equipment and more and more specialized equipment for all kinds of tasks involving physical labor because generally these things are easy to replace with simple mechanical devices and power.
China is currently in a recession. Maybe the growth is 6% or 3% or even less than 0%. Some part of this may be related to the fact that the kinds of labor the Chinese workers were replacing of making clothes, manufacturing small articles and components is more and more requiring the use of robots. The problem is that the robots are getting better but also that the materials that are being worked on by these robots are sometimes so small that it is impractical for humans to be involved anymore. As a result more and more things are being built by robots. If they can’t be manufactured by robots they may be 3d printed. 3d printers for anything from titanium to food products exist today. You can 3d print glass and practically anything replacing and allowing mass customization easily.
What about services jobs?
As these jobs were replaced by machines we have grown our services side of our economy. This has allowed us to grow the economy, grow the number of jobs even though large amounts of the previous jobs people did are not done anymore by humans. However, there is already evidence that computers and devices are starting to take over more and more labor intensive and even service type jobs. I begin to wonder how does the workforce and the jobs people do change. The main jobs at threat from this technology are jobs that don’t involve creativity or don’t involve having to switch from one specialty of knowledge to another or combine specialties of knowledge.
In the home we see the rise of things such as vaccuum robots, rug cleaning robots, lawn robots and I believe before long pick-up robots and other robots. The evolution of these robots is straightforward with current technology. We have developed robots that have smoother movement and are able to traverse paths easily. We have sensors to make it easier for them to see the world around them and to operate in the delicate manner needed and we have the automation and AI to enable them to recognize language, visual cues and objects so that they can interact with us and the world. It is not long before we have robots for the home that can do things such as pick up the house, clean surfaces, do maintenance type activities around the home.
When we look at the world of the service industry many jobs are rote, for instance, waiters, taxi or limo drivers, bank tellers and many other jobs may be easily replaceable with computers and robots.
We are hearing that companies are expecting driverless vehicles within 10 years or less.
Robot Doctors, Lawyers, Computer Programmers
However, much more serious inroads by computers are just starting to be felt. Doctors have long wondered if their jobs could be replaced.
A doctor has a certain intuition about what is wrong with you and how to treat you but in fact their advice is limited in most cases by their knowledge. We don’t want doctors to be creative in general. We want them to consult a vast store of information about symptoms and maladies to diagnose you and use proven techniques to fix you. All of this is well known information but making a computer that could recognize your malady from symptoms and then choosing a path of treatment has eluded computers for some time. It’s not like we haven’t tried.
Well, with the new recognition technology used in other forms of AI there is hope that they computers will be able to do much better at recognizing and doing more complicated things. Like becoming the best GO player in the world there is no reason to think that computers can’t fairly soon become extremely proficient diagnosticians. This won’t mean at all the end of doctors. Doctors use a number of skills and take into account many other factors in making decisions and prescribing treatment and it is unlikely we will allow a computer to make a life or death decision without a human overseeing the decision. Doctors will still be required in many cases to perform procedures. Some procedures may be automated but we will probably have doctors for some time but their jobs may change tremendously and we may need far fewer of them.
The same is true for many professions. Lawyers frequently are used for lots of tasks that are very rote. eDiscovery and common legal documents don’t take a creative mind to produce. Sure we will always need lawyers to argue the cases and to make the judgement calls but much of what we call lawyers today may be automated.
Computer programmers. I have considered this one of the last refuges from AI. The whole point of computer programming is creative generally although there are many rote tasks. The industry is in a constant state of improvement in reuse and generalization that results in greater and greater productivity but our need for programming is related to the takeover of computers and the very nature is to move to ever more creative aspects as the central part.
I believe that computer programming will actually be a basic skill that almost everyone needs to know because our life will be more and more about computers around us. As a result it will become more and more impossible to live without computer skills. No doubt the computer languages will become simpler and simpler almost to the point of natural language. However, there will always be a need for creativity to eek out the last percentage of efficiency and scale. I believe this is one of the safest skills.
Can we replace all these jobs?
As more and more jobs are eliminated or changed from our society people will have to find jobs. It’s important to note that in many of these jobs people may find the robots are simply an aid not a replacement or allow the profession increased productivity. People will be more and more employed in creative tasks and in jobs where human contact is required. I doubt even with robots that the prostitution profession is in jeopardy even with the threat or benefit of sex robots. People will also want humans in other places.
I suspect we will be able to create far more diversity in objects and manufacturing. In the science fiction movie Dune the future is filled with ornate art and luxurious complexity more associated with the Renaissance than Minimalist design associated with what we thought would be the future.
Up to now a large part of computers and robots has been about creating common designs and minimal complexity but robots have the ability to be programmed to customize easily. I suspect that more and more we will find help with creative aspects of everything and more and more customization. I don’t know if this can be enough jobs to replace the jobs lost to robots and AI or if those jobs will pay better than the rote jobs of the past. I suspect that there will be more and more an elite of people who are doing truly creative work that is above the ability of ANI robots and AI who get paid incredible amounts of money for their brilliance and more and more people who can’t meet the skills of these automated devices and algorithms and they will be relegated to more tasks we would call on to require the human touch. Maybe more massage therapists, more fashion designers, clothing pickers for people or individual designers for everything. Maybe more people in entertainment business or in news and reporting.
There are a lot of possibilities in which we have a stable world in which the number of jobs for these new things and growing occupations is able to replace all the jobs lost elsewhere and command better income. However, there are a lot of situations and worries that there will be a growing divide between the super smart and the average person. Who knows where each of us fits in this continuum and how our lives or children’s lives will change.
There will be increasing pressure to make sure every child is smarter and more creative. Increasing ability to modify our genes will likely put pressure to find a way to make everyone smarter. Possibly people will become smarter naturally. It is estimated the IQ of people generally has risen 20 points in the last 50 years or so in America. This is surprising because people I knew thought that humans were declining in intelligence generally. Some people even argued essentially a eugenics idea that smart people had fewer kids and dumb people had more kids so the population would get dumber with time. However, that hasn’t happened. Quite the opposite. I don’t think anyone knows why this has happened. Does the brain have more plasticity than we imagine and we will rise to the occasion naturally. In any case I believe this is going to cause incredible societal stress over the next 5 decades. We are really at just the beginning of all the change in labor markets and technology, robotic and AI integration. I worry about the future.
Why do I worry?
Humans in the television series “Star Trek” in the future behaved according to the Roddenberry doctrine. This optimistic view of man said that if we meet the needs of people they will become less antagonistic, more rational, less violent. The feeling of Roddenberry was that people would spend their copious time in activities of general positive consequence to humanity and to themselves. Certainly many people might do this but there is a dark side to the human mind, character or personality that is undeniable. Having an alcoholic father myself and other people I’ve known I know that people can fall prey when they are idle (or even NOT idle) to overwhelmingly negative behaviors to society and themselves.
If most people were given an income and free time they would undoubtedly spend it in a positive manner but there can be no doubt that a subgroup that would spend their free time and money in pursuits of extremely negative consequence. I have no answers to these problems. I have no idea how Gene Roddenberry imagined these things would go away. He apparently had a profound faith in human beings being basically good. Maybe we discover the causes of these negative behaviors and we are able to filter or cure but this has tremendous philosophical, moral and practical problems.
What about AGI and ASI?
This article is really about the implications of ANI on society and that all the issues of employment and worry about our future can happen with just the ANI technology we have or are close to having. The future may bring about AGI or ASI. These complicate the future because if AGI is possible then we will have an alternate species competing for even the most prized jobs. We will have to face the idea we have created another sentient life and that this life may have rights. If we develop ASI which is super intelligence beyond humans we have more problems since the new creatures will be superior, how will we control them and can we control them and is the human race extinct? Does human intelligence mean conflict and inevitable extinction of one species? How would we live in a world with different levels of intelligence or different species of equivalent humans? The history of the human race isn’t too positive on this.
ANI is challenging enough and is upon us. We need to make decisions how to build a world with these technologies, genetic technologies and other new technologies that allow us to dramatically change the condition of the human race in the next 50 years. I am staggered personally at the challenges we face but also optimistic that like the past 300 years since the industrial revolution and the rise of the scientific method, democracy etc that man has benefited from each new discovery and improvement. During these improvements there have always been nay-sayers and they’ve always been wrong but we don’t know if that record of continuous human improvement will continue or if these technologies will ultimately lead to a nightmare future. It almost makes one want to believe in god. Can somebody help us please navigate these waters ahead? Can some superior race with extremely beneficial motives help us or do we have to do this all alone?
Other Resources you may want to read: