Pandora's Box: Hope always remains

     Most everyone has heard of Pandora’s Box (it should really be referred to as Pandora’s Jar) but the details are not always that clear. In Greek Mythology (and myths are those stories that help to explain the world to the surrounding society), Pandora was the first woman created by the gods. They gave her lots of positive attributes, each according to the strength and power of each individual god. They also gave her a sealed jar which was to be kept unopened (does this sound a bit like an apple in a garden?). Since curiosity is one of the prime emotions of humanity, that forbiddance was only an enticement to see what was inside the jar. When she opened up the jar, the evils of humankind leapt out into the world … but hope remained.

     The story continues to carry force just as humanity continues to have imagination and curiosity as a basic aspect of our personalities (not for everyone, but for many). In particular, humans seem to be fascinated with technology. Or, at least, advances in technology are the ones most honored by society. There are certainly those who work on philosophical, spiritual, meditative, or interrelationships that could make even more of an effect upon society if society were open to their movement and growth.

     And there we have the struggle. Oh, it is not a precise “them versus us” type of struggle. But, since technological progress can often be inserted within the economy and, thus, “valued” in monetary units it is most often valued above progress in behavior and humanity. And that is sad.

     Almost every (perhaps every) technological advance can be used for the improvement of humankind or for the advancement of a small subset of humankind. Going from the Stone Age to the Bronze Age allowed for better cooking utensils, better diet, and general tools to be used for art and construction. It also allowed for more efficient weapons. The same holds for the movement from the Bronze Age to the Iron Age.

     Perhaps, if the evolution of technology went hand-in-hand with the evolution of society, life would always improve with each step (forward or backward? A matter of opinion). I don’t know if that will ever be known as it has yet to happen, in my limited knowledge.

     During the age of the Luddites in the early 1800s, the technological step was towards automation. The displacement of skilled labor was a huge strain upon the populace and, since the general society did not recognize, and allow for, this change it was a matter of great concern to many — many who did not have monetary, or societal, influence. In such a situation, groups of people often fall back to violence even though they know that such actions will not have a long-term positive effect.

     Two hundred years later, we have yet to fully absorb the changes to society brought about from automation. And now we have a useful generative AI inserted into our society. Understood well by general society — no. Understood, and utilized optimally by general industry — no. Societal effects upon the general population researched and worked with for a gradual, and less painful, shift in society — no.

     But we will continue to move forward. The technology cannot be “put back into the bottle”. We could learn from the past and research the general societal and economic effects from such technology but will we? Ah, that is up to all of us — and that does not give us good odds on the results.

     So, here comes AI “out of the bottle”. It is causing turbulence and will cause more (just how much is yet unknown) along with economic displacement. As is true of other technologies, there are “two edges”. But, as is also true of prior technologies, there has been no concerted effort to prepare for the shift in society, jobs, and integration with other factors which will be caused by the introduction of generative AI (and, it shall be even more difficult with more generalized AI in the future). Perhaps, somewhere, there exists an authoritative group that deals with serious interactions between technology and people — but, instead, we keep getting political and profit-based incursions into public decisions.

     We can do much, much, better — and we need to do such.

Thanks for reading Ideas & Interpretations! This post is public so feel free to share it.

Ideas & Interpretations is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Displacement: Physical and Economic

     Life is change. It can be slow, fast, or abrupt — but it happens. Our planet has gone through changes throughout its history. Asteroids have hit the planet. Volcanoes have caused short-lived, but deadly, global winters for the planet. We have had Ice Ages and receding periods. Continents have split apart and merged.

     We are currently in a global warming period emerging faster than has happened in any other recorded period. This is based on comparisons with historical geologic and ice core readings. It is due to increases of greenhouse gases which prevent heat from escaping the earth’s atmosphere and which causes a change in the balance of energy. The oceans are warming and that gives a greater heat/energy source for creation of larger, more frequent, drastic weather events such as droughts, floods, storms, hurricanes and — yes — even blizzards. Almost all scientists agree that humans have greatly contributed to those greenhouse gases. It is not as certain as to whether it still would have happened without humans — but, if so, not nearly as quickly.

     But it IS happening and it’s happening quickly. The changes in weather patterns are forcing changes in crops grown in different areas. The rise in sea level threatens coastlines and many island nations, with much of their area to be below estimated future sea levels, are making plans for migration (or emigration) of much of their population. Larger, non-island, countries will have to deal with coastlines receding inland with established coastal development going under the water. Other large-scale displacements of people will happen because of changes in availability in amounts of potable water. More energy in the system can power many different events and 100-year-floods become 5-year-floods.

     People are not displaced solely because of physical reasons. The tides of change in the economic arena can move people even more than an earthquake.

     The name “Luddite” (named for Ned Ludd who destroyed some mechanical equipment in 1779 out of frustration with changes happening) can be found in literature around 1811. In this region of time, the “Industrial Revolution” was taking place. What once needed ten people to do the manual labor for a task could, with machines, be done with one person (trained differently and with a different set of daily duties). What happened to the other nine? They were unemployed — and, since there were a lot of instances of this happening in the textile industry, there were a LOT of unemployed people without any type of modern “safety net”. They had been displaced by the new technology and machinery.

     Many of the newly unemployed workers had done the same type of work all of their lives. The companies that they previously worked for had no legal obligation for generosity and compassion to help the unemployed workers — nor did the religious leaders of the time urge such!

     Over the years, technology has replaced technology. As Danny DeVito says in “Other People’s Money” (not the exact words) — “I’ll bet that the last company that made buggy whips made the absolutely most wonderful buggy whips ever”. But the buggy and coach have traveled over the hills unlikely to return.

     Over the past 50 years, we have had white-collar labor practices be replaced by programs and computers — and blue-collar positions replaced by automation. It has caused discomfort and displacement but manageable. Now, the process continues to accelerate. Though we have yet to see just how effective, and useful, current AI methods are within society, if they come even close to the use as hoped for by businesses, we could displace 10% to 20% of the labor force. These are not people who have yet to find a job. These are people for whom no job exists. What happens with them? Desperate people do desperate things. A universal base income might keep them alive but there are so many other societal aspects that blend into that and other possibilities.

     While this is something that I have said, and will say, on a disgustingly regular basis; we have problems to approach and work towards solving — but the first overriding problem is getting people to talk to, and work with, each other to work to solve them.

     But we can do it if we want to do such.

Ideas & Interpretations is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

AI Caution: AI does have limitations to be aware of

     I took an AI course at the University of Idaho in the fall of 1976. Forty-nine years later they are releasing products that have that name associated with them. Is this the same AI as was talked about in 1976? No, not really. If you search for what kinds of AI exist, you will get various lists according to capability, functionality, and techniques/approach. The current AI implementations are at the beginning levels — often called “Narrow”.

     The course that I took in 1976 was more a matter of showing early investigation of programs such as Eliza and early approaches to Machine Learning (ML). Memory capacity, as well as CPU capabilities, were very, very small compared to that of today — and that limited implementations of various approaches. As is true of all computer programs, the main advantage of the computer doing it is that it can do things very, very, fast. Today’s computing power can achieve results only dreamed of in 1976. (And, if the reality is desired, it takes a lot of electricity and hardware to provide the capabilities even today.)

GIGO

     Current Narrow AI implementations such as ChatGPT, Siri, Google maps, or facial recognition are focused on a single type of task. Even the task of answering questions (ChatGPT, etc.) is a single focus. Responses are made according to the data that the program has had access to — and the training (yes/correct, no/incorrect) about appropriate responses. Two problems (among many) arise — are the data that are examined valid (do they meet definitions of facts and correctness) and is there bias (usually according to the trainer) — if they are given proper data but the trainer tells them to reject the proper data in favor of dubious, or clearly incorrect, data?

     Whether it is because of faulty data sources or deliberate subversion by the trainer, these are instances of “Garbage In Garbage Out” (GIGO). This is an old term used by early programmers but it is still valid today. If bad data are part (or all) of the input then the output cannot be trusted. A search response collator (such as ChatGPT) must be treated the same as individual searches which bring back various “hits” at different websites. In other words, even responses from “AI” bots need to be fact-checked.

     Even though you should not blindly trust the output, the use of a Narrow AI can be very useful. It can be used to very quickly gather, compare, and present results similar to the results if you had done a search (or multiple searches), read through the contents, and correlated the contents. The differences are that the Narrow AI can do it much faster but, since it does not have true creative/interpolating intelligence, it does have the caveats mentioned above.

Creativity

     When a Narrow AI presents you with a result in the form that you might submit somewhere or directly use, recognize that this is an “average” of many sets of example data. Although the definitions of such things as “average” are a bit vague — you will be presented with results that are somewhat the average of many possible results. It is very unlikely to be of the very highest quality but is likely to be adequate for your purposes. Personally, unless you have absolutely no skills in the area of the query, I would suggest using the AI result only as a starting template. Change, emphasize, and make it your own.

Permissions

     When a Narrow AI is trained on data, it just grabs data from wherever it has access. When presenting results, the sources are fully hidden from the user. Unlike what would happen if you did the same research, it is difficult (perhaps impossible) to know how much came from this and how much came from that. There is no explicit cautioning about taking the source into consideration and respecting intellectual property rights. This is a legal issue that needs to be pursued (soon) by various governments and legal bodies.

Timeliness

     When you do a search through the Internet, you have the option of age of responses. However, one of those options is to get only the very latest (say — past month) data. When you use a collating “AI” search bot, you do not know just how old the data are. During the first beta testing periods of the various collating search bots, they had an explicit warning that data was obtained only from sources older than a specific date — current data was not included. Lately, such warnings have disappeared but that does not mean that the latest data is incorporated. Given the way that these programs must be trained, it is very unlikely that the latest data are used.

Summary

     Use of an AI program can be beneficial. It can help you create a good, average, example of your desired result. It can speed up searches and correlations of large amounts of data. Results will not be as good as that from expert human intelligence. And there are potential permissions conflicts. Do not treat AI collating bot results as firm, objective, truth — treat all results the same as you would of any other search of data throughout the Internet — with care and a need for fact-checking of potentially subjective results.

Ideas & Interpretations is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Ethics: Always in process

 

     There is a lot of talk about ethics currently -- especially in relation to developments in AI. According to the dictionary, ethics are the moral principles that govern behavior. Note there is nothing about "universal" or "unchanging" in the definition. Ethics are based within a specific society and their rules of behavior. Societies vary a lot in their viewpoints on many issues -- sex, gender roles, religion, death, expected behaviors, taboos, and so forth. A society that expects resurrection of the individual will look at death a lot differently from a society that expects only one life (whether they believe in an "afterlife" or not).

     Ethics also change along with societal norms. Fifty years ago, there were a set of behaviors that were normal, and expected, from the members of the community. Now, in this present time, some of those behaviors are no longer acceptable. The ethics relating to current societal norms may reasonably be applied to behaviors that are happening now. They can also be applied to past societies, writings, and other memorabilia but ONLY as if one is using a microscope. They can be examined and the difference between the ethics of that time and the ethics of the current time can be looked at as documentation of change. But the past behavior WAS acceptable at the time that it happened because the rule set, the ethics of the time, was different. Each period of time and each distinct society has its own rule books that should not be applied to judge behavior within other times/societies.

     The above is true -- but only in the abstract. In real life, everyone believes that their current ethics are THE correct ethics and ethics that differ are WRONG. People with different ethics -- whether of the past or of a different society or culture -- are BAD. This would be true in the opposite direction of course -- people of the past would consider OUR behavior to be BAD when judged according to the ethics of the past. This almost always leads to conflict and even to wars.

     A cultural anthropologist must always be very careful upon entering into a different culture/society because they are there to examine, analyze, and document. Since the new ethics are not known, it is easy for them to violate those in some manner which would make their work much more difficult or even impossible.

     We have been talking about the flux of ethics. Within the current time and current society there is a set of ethics rules that applies to everyone within that time/society. If they do not follow them they have to face social, and possibly legal or martial,  consequences. This is what is typically called an ethics problem.

     AI can face issues similar to that of human individuals. One is the ownership of intellectual property. For humans, taking such is called plagiarism. Generative AI systems need to be "trained" by giving it access to much information -- not all of which is legal for general use. In that way, AI may be considered to be plagiarizing or even stealing from others within the same subject matter? It may also affect the value of those systems, or people's valuable intellectual property, from which it absorbs information? I put question marks because those are part of the questions posed about defining ethics for AI systems.

     Humans can (and do -- much too often) lie. They do such deliberately (called with malice) and accidentally (by not verifying information before passing it along). If AI is trained with bad information (deliberately bad or non-verified information) then IT will use that bad information and pass it along as "good". For humans, this is called slander or libel -- but often not subjected to legal ramifications if done within social media. What is such when done by AI? Can this distortion of reality, and worldview, be considered a dangerous crime by AI systems? Another area of ethics in connection with AI systems.

     Humans can, and do, commit crimes as defined by the legal systems of their society. AI systems can be trained to do such faster, and possibly less noticeably, than that which is done by humans. AI systems can be trained to "phish", steal private (supposedly inaccessible) data, forge accounts, and other non-physical actions. If given access to "waldos" (physical systems capable of being controlled online) then they can even do physical crimes. What is the legal aspects, and ethics, of using AI systems for such actions?

     Humans can, and do, violate ethics systems. AI can be trained, or designed, to do similar violations and be done faster, and more effectively, than humans. These are the areas that badly need to be addressed before the problems that will occur become too unwieldy to deal with.

Frankenstein's Monster: AI's shadow

 

     Alan Turing, in 1950, released a paper called "Computing Machinery and Intelligence" while at the University of Manchester. The main gist of the paper was that if the output of a computer could not be distinguished from that of a human being then it must be considered intelligent. This has been a goal of those working on AI for a long time and it is the primary point of general acclamation that we "have achieved AI". Various AI applications can present artwork, prose, dialogs, and other such things that cannot easily be distinguished from that created by a human. In a "blind test" (where the people judging have no pre-knowledge of which output was done by a computer) it would pass judgement.

     No one is saying that AI has reached the goal of achieving human thought. AI is trained (as is true of humans) on lots of input -- lots of data from lots of sources and, iteratively, told what is right or wrong and "learns" how to do things, respond to things, and so forth. Which leads us to two of the current problems with AI. These are intellectual property laws and lack of discrimination.

     Intellectual property conflicts are the most straight-forward. In order to train, lots and lots of information must be fed in. While some of that information is data obtained from various public sources, other information is from private, proprietary, sources and other information that has been created by human minds and may be part of their livelihood as well as their career growth and reputation. That "brushstroke" done by a generative AI may be copied from an artist's work. That "narrative work of imagination" may be borrowing from the insights of dozens of writers who, justifiably, do not want their copyrighted material used without their permission. AI generated code may (and almost definitely will) make use of code that was created both for freeware as well as company proprietary code (that probably was not authorized to even be accessible via the Internet).

     Currently, an AI information gathering product can present bad information. As an AI model is being taught, it is given access to a lot of information. Hopefully, most of this information with be factual and truthful. But some of it will be misinformation (lies), disinformation (logically faulty or irrational), or declared works of fiction. Some of these can be eliminated by human monitors as the model is trained but this just leaves it in the hands of those humans.

     Humans do not have a good track record in doing the research needed to determine truthfulness of information -- they cannot be relied upon to do such for an AI model. It is conceivable that an AI model might be programmed with algorithms that will allow it to compare information, reliability of sources, and logical inconsistencies. They even would have an advantage since they do not (currently) have emotions. But AI models are NOT at that point now -- and it will depend on the accuracy of any such algorithms.

     Isaac Asimov, considered one of the best of the "classic" science fiction authors, used as part of his robot-focused novels an idea called "The Three Laws of Robotics". I'll let people investigate the evolution of the laws, direct quoting of the laws, and the interpretations as discussed within the various stories produced. Let's just say these "laws" are meant to be a "leash" to prevent robots (or AI) from hurting people. And, though the laws work pretty well in the stories, there are almost always loopholes in any "law" and the three laws are no exception.

     But current AI has no such thing as the three laws -- and no practical way in sight to create such laws and to enforce that they will be part of any, and all, AI products. While AI is monitored, and final decisions are made by humans, then the moral (and legal) responsibility for any problems will be in the hands of humans. If unmonitored and autonomous, it is unknown as to where the responsibility would lie. There are presently court cases struggling to delineate such matters in the relatively straight-forward arena of self-driving cars.

     Computers, and programs, are NOT "smart". Compared to most humans, computers can be considered rather poor in intellect. They can only do very, very simple things. Arithmetic, comparisons, actions based on values, and such are the very limited set of actions that they can perform. Computer programs eventually are split apart into these very simple "machine" instructions. BUT, computers can do these simple instructions very, very fast (and getting faster every year). This gives the illusion of great ability by the computer.

     Humans make mistakes. Computers (or their hardware and software) can make mistakes. But computers can make many more mistakes, much faster, than humans because their basic strength is speed. Plus, the more humans rely upon computers (and their hardware/software) the greater the difficulty of correcting errors because many people erroneously think that computers "cannot make mistakes". Even worse, humans are often not allowed the ability to override the decisions that come out of computer programs.

     This is particularly relevant in many of the areas of AI. People do a poor job of facial recognition. Computers can do a better job but not a perfect job. If drones are given a locate (and, perhaps, violent action) instruction autonomously then they will likely sometimes pick the wrong person. The situation is, of course, even worse if computers get to make the decisions about all-out war (such as in "WarGames", "Terminator", or many other apocalyptic films). Politicians are quite at ease in having "collateral damage" but if you, or someone you love, is part of that collateral damage then you would probably not feel as calm about it. If facial recognition is done as part of security then "doppelgangers" or others who look similar to those on no-fly lists (or other lists) will forever be fighting for their right to exist.

     Computers, their programs, and AI can be of great benefit to humans. But they should always be subordinate to humans because humans are the ones that suffer from any mistakes made. Recognition of the fallibility of computers is very important and self-regulation within AI programs is essential to reduce the escalation of errors.

The ever changing work week; history and technological change

 

     Once upon a time, long long ago, people did not divide their time between work and non-work. People spent the time needed to survive and did whatever they wanted the rest of the time. In certain "tropical paradises" with ample food and few dangers, most of the time was spent doing sports, enjoying flowers or doing crafts, building monuments, or whatever. In less amenable environments, most of the time was spent trying to survive and the only non-work time was during winter months when activity was severely restricted (but, often, the time was still allocated to doing activities to prepare for the warmer periods).

     As humans developed technology, including agricultural, animal husbandry, and environmental adaptive technologies, the less hospitable areas had less time required for survival. Society moved in two directions to make use of this non-work time. One direction was social stratification -- some had lots of leisure while others lost their leisure to perform labor for the higher social layers. The other direction was religious and familial development which provided more structured activities for use of non-survival time.

     Societies which had developed in the less stressful areas did not shift as much -- continuing to spend most of their time at leisure activities. In the other areas, a dynamic developed between the two directions of the shift of people's time with some wanting to do less and less for themselves and the others wanting more time to develop leisure activities. This dynamic continues unto today.

     But technology continued to develop and this push towards reallocation of survival labor time to non-work changed. This was buffered, in part, by continued population increases. The more people, with the greater strain on resources, the more efforts needed to survive -- but technology changes have continued to move somewhat ahead of those needs.

     In some areas, bolstered by religious mandates, the working people started having times, and days, where mandated activities could be forgotten. Children started to be seen as developing people rather than just another labor resource. Days were allocated for worship. Lastly, the work day was shortened to allow non-work time before and after the work day.

     In our present time, between automation and AI, we find our society entering a new period when the amount of work hours needed is becoming a smaller, and smaller, part of the total number of hours available to be produced by people. And, with automation and AI, the training and abilities needed by people who do work becomes more difficult to achieve.

     This leaves us with many hours not needed to be worked. It is no longer a survival situation -- it is a political and economic problem. If we reduce population then all of our production, and economic flow, will be at overcapacity (more than we can use). If we do NOT reduce population then there is not enough work for everyone. One possibility is to have each working person work fewer hours. Shorter days or fewer days? With commuting, fewer days means less overhead and more efficient labor movement.

     There is probably no universally "correct" solution but, unless civilization does a "reset" due to societal collapse, we need to face the problem of not enough hours of work for the number of people we have. Many object strongly to people not working but, if that is to not be the case, there must be better distribution (and better education) to spread the number of work hours among the many people.

The Turing Test: Not enough?

 

     Of late, there have been declarations of Artificial Intelligence (AI) becoming sentient. And even louder protestations that it isn't true. Within computer science, there has been an assumption that this can be evaluated by use of what is called "The Turing Test". Created by Alan Turing, a British mathematician and considered to be a founding person in AI, in 1950 -- it was initially called an imitation test. The idea was that if you were unable to distinguish the conversation of a program from the conversation of a person the program could be considered sentient.

     This was in 1950. People conversed on a regular basis in 1950. The criteria of being able to distinguish a person from a program, by use of analysis of conversation, seemed reasonable. Forward the clock to 2022 and that criteria no longer seems to be sufficient. For those who say that AI has reached the criteria set out by Alan Turing -- they are probably correct. There are programs that speak language BETTER than an average person in that language. According to the Turing test, they are sapient. But, according to our current experiences of programs, machine learning, and AI -- it seems that it would be hard to consider them to have had a passing grade for sentiency.

     Which leaves us to the question, once again, of what is sentience? Some of the recent declarations indicate the appearance of emotions, desires. Are these what distinguish sentiency from mechanical responses? Of course, one can move to the religious, spiritual, and theological and say that what we are searching for is a soul. But, if it is a matter of a soul then we are in even worse shape for evaluation. There is no consensus as to what a soul is. If it does exist, where does it reside? If it does exist, is it immortal or connected to something associated with a living mind and only in existence as long as that mind and body exist?

     Science fiction and fantasy -- those genres of "what if" -- have approached such a subject for many decades. But general agreement, even less the matter of consensus, is nowhere to be found.

     If we use the criteria of a desire for survival and a desire for progeny then the state of AI has not been achieved. But, if those are to be criteria, should a goal of reaching such be striven for? Mary Shelley's Frankenstein reflects many matters: human hubris, the difficulty of accepting others, fear of the stranger (xenophobia). But, as seen in the dystopia of the Terminator series, a desire for survival and progeny coordinated with a superiority in speed and execution may prove to be an ultimate competitor to the human species.

     So, let's move back to the start. If we use the Turing Test, alone, as a criterion for sentiency -- we are probably there. If that is not sufficient, what qualities should be measured -- and how should they be measured? And, in some situations, perhaps creating that which meets such criteria may be something that would be a mistake to achieve?

AI and the "3 Laws of Robotics" : Could it be done?

     For those of us who crawled under the covers with a flashlight as we were growing up, Isaac Asimov will probably always be the first thing we think of when the word "robot" turns up. Of course, in Asimov's world robots were always humanoid in appearance. In today's world, with robots becoming more and more integrated into society, we recognize that a human-appearing robot will fulfill only specialized niches,

     The robot (from the Czech word, robota, for "forced labor") only earns that definition if, once given a directive, it can work toward a variety of goals without further instruction. This ability is based on Artificial Intelligence (AI) (though I doubt there is a consensus on the definition of AI). The more, and larger variety, of autonomous actions that can be performed the "higher" the level of AI. Some (such as Alan Turing) believe that AI has only been achieved if one cannot distinguish between the "artificial" responder and a human one.

     Humans have the distinction of being able to do a huge variety of activities, including activities that have never been done before. So, for AI to approach human capabilities, the mechanism would need to be programmed to do them. Hundreds and thousands of tasks. Often such tasks are truly composed of "subtasks" such that those thousand tasks may be made up of, perhaps, 75 subtasks put together in different specific sequences.

     This is rather a lot of work, however, and it would be so much "easier" if the robot could determine how to do new tasks by itself. If a robot is pre-programmed, it can only do what it has precisely been told how to do. "Scratch the right side of the nose" is a different task from "Scratch the left side of the nose". (Sensors could be installed that activate a signal called "itch" and then the program could say "scratch the place that itches".) This self-programming is often called machine learning, which is considered to be a subset of AI.

     Many people are afraid of robots taking over from humans. Whether they develop a contempt of humans -- as within the "Terminator" series or being used as supersoldiers to control, hurt, or kill humans. In order to help people have a more compassionate and trusting view of robots, Isaac Asimov invented the 3 Laws of Robotics which are as follows:

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2. A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

This is a hierarchy of importance: human life, following orders, self-preservation. For humans, self-preservation may sometimes take the top position -- the idea being, once again, that human life is of more importance than robotic life.

     We will deal only with the "First Law" in this blog. This breaks into two categories -- active harm and lack of prevention of harm (when known and possible). For pre-programmed robots, it is easy to avoid the first -- just don't program that task. For self-programmed robots, there needs to be an "override" program to prevent action if it will cause harm.

     The second, prevention of harm, is much, much, harder as it requires "judgement" -- deciding how dangerous a situation is and what various contingencies might arise that put people into danger. This is the situation currently facing those who are trying to program self-driving cars. The problem is similar between pre-programmed and self-programmed robots/mechanisms. For pre-programmed robots, all of the contingencies must be thought of -- good luck, but it isn't likely to be possible. It is possible to envision a self-programmed robot to "learn" enough to be as good as a human in their judgement -- but we currently don't know how to do that. Maybe we will learn.

     The "other hand" to this, however, is that implementation of the "First Law" is purposeful. It must be done deliberately. There are too many people, and too many groups, that currently have people deliberately hurting other people. It is not reasonable to think that usage of robots will be any better than usage of people. This leads us to the world of the "Terminator", programmed face-recognition assault drones, supersoldiers and all of the nightmares of technologic fear.

     In Asimov's world, "U.S. Robotics" had a monopoly on building robots and was a benevolent entity that insisted on the 3 laws being in each "positronic" brain for their robots. In our world, there is no effective control of technology in a preventative mode (we can make laws to penalize use after the face). I cannot imagine how we can prevent some of this harm from occurring but there are a LOT of very creative, imaginative, intelligent people out there. Please start working on the problem while it is still theoretical.

Controlled falling: how to teach a robot to walk

     Every part of growing up is a miracle in its own way. However, if you happen to be an engineer or a computer scientist, you may find yourself looking at your child in a somewhat different way than most parents. Every act is a matter of "how did they do that?" Or, a matter of "I didn't know they couldn't do that originally".
     Learning to walk is a gradual process. The first part is a matter of figuring out just how to control those wonderful muscles on purpose. For fortunate babies, they have a working nervous system and all of the appropriate muscles are there but that doesn't mean that they pop out into the world ready to do a 100-yard dash. Think of a control room with hundreds, or thousands, of unlabelled switches -- each of which cause a muscle to respond in some way. How do we use an electrical switch box which has lots of unlabelled switches? Try them out and see what they do. (And then, perhaps, label them after we notice their effects.) For a robot, this is a bit simpler as there is a specific control register (or bit within a register) that causes a specific servo-motor to work.
     Now that the child knows what muscle connects to each impulse (and I am not going to try to pretend that I know just how this really takes effect), she (or he) has to practice. This may entail kicking dad in the face a few times and laughing or hitting brother in the nose. Strength is developed as the muscles are exercised. And a special sense (not always fully present in autistic children and others) called "proprioception" starts to be better known. Proprioception is also sometimes known as "body sense" or "kinesthetic awareness". No matter what you want to call it -- it allows us to know just where our body parts are. Is my finger extended? Is my leg bent? This is important if we want to apply the right muscle at the right time.
     For a robot, this has to be done in different ways (although, once again, I do not claim to know just how body sense works within a human). One dominant method is to keep track of relative position. This works like the cursor on a screen -- when the system is powered on, a specific point is considered "home" position and the cursor is moved relative to that position. The same can be done with any servo-mechanism between the limits of its movement. However, it must start at a known location and there cannot be any exterior limit on the movement (which would cause a need for recalibration). Other methods are possible but require more active sensors (and, thus, are more expensive).
     Two more requirements exist for easy movement. These are the ability to know how hard a muscle is pushing against something (the floor, for example) and how fast it is moving. The human nervous system makes use of tactile feedback to determine how hard the muscle is straining and the body sense to know how fast it is going. With a robot, a feedback loop using torque measurement may allow the robotic arm to hold an egg -- or to crush it. Speed is determined by the rate of change of movement -- how fast position changes versus and internal clock.
     With these four aspects -- ability to move, knowing where the parts are, knowledge of amount of force, and knowledge of speed -- coordinated movement is possible. Early programming of robots tried to imitate the specific movements of human muscles within their ranges of motion. It is possible to do it this way provided that there is complete control of the environment. Nothing in the wrong place, no unexpected alterations in the footing or the locations of other relevant objects. Consider a factory line with fully repetitious movement and behaviors (until a part sticks or parts run out or a dog runs into the factory ... or) and one can relatively easily see a robot taking over the factory job. In fact, many of the jobs taken over so far have been of this nature. 100% replacement is not possible because of the many exceptions that can take place and which requires more flexibility to handle -- but a considerable reduction in human staff is possible.
     But we were talking about walking weren't we? Could we use the same methodical programming to teach a robot to walk? Barely possible but, once again, only within a highly controlled environment.
     Imagine that child learning to walk. They stretch. They pull. They start becoming caterpillars on the carpet while they both strengthen and practice their muscles. Finally, they pull themselves up. And fall down. And go up. And fall down. Then they are able to stay standing up -- but hanging on. Then they let go. And fall down. And so on.
     This is a type of programming -- but not "linear" programming. This is not "do A, followed by B and then C". It isn't even exception-handling programming "do A, followed by B, then D if condition C else do E". This is neural programming. Sequences are attempted and then, based on results, discarded or modified or increased. A goal has been set and if enough sequences are tried then, at some point, success will be reached.
     Note that a new item has now been added -- a goal. In order to have a goal there must be a way to determine if you have reached that goal. For a child that is emulating other people who are walking. For a robot, it is necessary to have goals that can be specifically quantified -- expressed as numbers -- against precise targets. For walking that might be obtaining a certain height, directional velocity, and stability. Note that balance, for a human, is obtained by the feedback from the inner ear. Tools, such as gyroscopes, are available to both help maintain, and recognize loss of, stability. Laser positioning devices can be used to indicate height. Global Positioning System (GPS) information can be used for large-scale movement for direction and a combination of position and speed tracking can be used for shorter distance velocity calculations. I am sure that other tools also exist.
     For a child, they see others walk -- and those others encourage them (and protect and guide) -- and they go through a seemingly never-ended process of trial and error. They train parts of their brain and nervous system such that the thought "walk" indicates a complex series of changes, movements, and activities. I shudder to think of trying to program that linearly.
     A robot can learn in the same manner but they have to have ALL of the correct tools -- servo-motors, proper range of motion, torque feedback, auto-recognition, or storage (with its likelihood of losing calibration), of movement, and so forth. As long as they have a goal against they can match their efforts, they can keep trying combinations until they succeed. However, there is a "secondary" aspect of this type of learning -- to keep the "winning" processes and discard the "losing" processes. Humans do this (in some way that I cannot explain) but robots have to do it also. In many ways this is even more difficult because it is unlikely that the next attempt will be EXACTLY like the one in which they previously "won".
     As a note, other types of activities can be approached in the same manner -- trial and error measured against a goal. But the less physical the more difficult the definition of the goal.
     #robotics,#AI,#NeuralProgramming

Artificial Intelligence: Beyond the Turing test

     In 1950, the British mathematician Alan Turing gave an answer to the question -- how can you tell if a machine is intelligent? His (paraphrased) response was "if you cannot tell the difference between a human answering questions and a machine answering questions then it has achieved intelligence". This Turing Test is not universally accepted but it is probably the most widely used foundation of answering the question of what is Artificial Intelligence (AI).
     Alan Turing's test was based on the idea of an interviewer and a responder. Someone asks a question and someone answers a question. This led to a series of experiments in computer programs that simulated (or imitated) "normal" human interviewer/questioner situations. It might be between a therapist and patient or doctor and patient or a student and professor/teacher. Naturally, there had to be a way to make it impossible to physically tell whether it was a machine or not. It also had, built into the test, the requirement of equivalent skill in understanding and speaking/responding in a human language.
     In today's world, computer programs have advanced beyond simple questions and answers. We have computer programs beating humans in Chess, and Go, (and other games). The Turing Test might not be considered to apply to these situations but many people would consider this a form of AI. We have computer program/systems that make use of pattern recognition to identify potential suspects or targets of drones. So far, the final decision is still made by humans but stories/films such as The Minority Report indicate a possibility of the machines making final decisions even about what might happen.
     That is the "line in the sand" for people thinking about AI. Who makes the final decisions? Is it a human (with all of her, or his, faults and experience) or a machine (who, at heart, is still the results of a programmer's abilities and recognition of exceptions)? Isaac Asimov, in his Three Laws of Robotics, had the AI programming include self-restraints as to what the program/robot could do, or could not do, without undergoing self-destruction.
     Speed and safety. The primary reason for computer programs is NOT that they can do things that humans cannot do; the primary reason is that they do things much, much, faster (and reproducibly). So, if you design an AI that handles the coordination and operation of a nuclear reactor, you want the program to be able to respond very quickly. Putting a human into the decision path slows everything down. Who has the final responsibility?
     The same question exists within the possibility of self-driving automobiles and trucks. It is likely that AI programs can already drive as well as an average driver -- assuming that all of their sensors work properly (they can detect objects and highway lines and sounds and bouncing balls and the cars and buildings around them, ...). Certainly, in another five or ten years, AI self-driving programs will be able to control a vehicle much safer (and more rationally -- no road rage potential) than humans. But they would be making the final decision.
     If a self-driving AI makes a mistake or a necessary decision that costs lives, who has the responsibility? The programmer? The company that built the vehicle? The owner of the vehicle? What happens if the self-driving car is involved in an accident with a human-driven car? Is there presumption of innocence on the part of the self-driving car?
     In all these cases, the program and machine are taking the place of the human. If you keep them "behind the curtain" there may be no way to identify whether they are human or machine. They PASS the Turing Test. But, when the curtain is removed, what is the final verdict? Who/what has the responsibility? Who/what makes the final decision?