Computing power, data storage and transmission speeds have all continually advanced from the moment counting tools were first employed to extend human mathematics beyond our ten fingers. Calculations were aided and our ability to do higher order math improved through the use of tally sticks thousands of years ago. The abacus was introduced in Babylonia nearly five thousand years ago. The slide rule, a major step forward in being able to complete highly complex mathematical problem-solving was invented by William Oughtred sometime between 1620 and 1630, shortly after the introduction of logarithmic numerical scaling.
The first mechanical computation machine was invented by Charles Babbage in 1822. The first electro-mechanical, automatic computation machine, now called a “computer”, was invented in 1941 by Konrad Zuse. In 1945, another significantly improved computer called ENIAC became fully operational, which used the same vacuum tube switches that Zuse employed. ENIAC weighed thirty tons. It was the invention of semi-conducting materials which allowed components to be made smaller, with less heat and faster operational speeds, that began the incredible rate of advancement and reduction in size of computers.
Science and technology continued to accelerate. Even the rate of acceleration of advances accelerated. Gordon Moore, the founder of Intel, coined what became known as Moore’s Law in 1965, which said that transistor density (and thus computing power) on a semiconductor device would double every two years, roughly. We are now reaching the end of what is physically possible in device density within the context of semiconductor materials which employ electron signaling to represent the ones and zeroes that comprise the foundation of all computer instructions, or “languages”. Quantum computing, which offers more information per electron due to the property of superposition of states of subatomic particles, has the potential to again accelerate the rate of advances in computing power per unit of device real estate.
In recent years the increased power and accuracy of our machines, coupled with the incredible volume of data available through decades of use of the “world wide web” on a massively expanded internet infrastructure, has resulted in the ability to compute and therefore identify complex patterns within gargantuan databases. Now we can use those patterns to create novel content. These programs (and they are indeed software tools which must be designed) which use the available data are called “large language models”. The languages referred to include visual, numerical, symbolic and abstract elements, as well as traditional word-based languages. The large language models are told to go looking for patterns and to be able to use the patterns to generate responses to queries, such as those posed in ChatGPT’s online tool. The responses can be both unique as well as repetitive of existing content. The responses can also be wrong.
The debate we are having is whether these new tools can be said to represent artificial intelligence. The startups looking for investor financing or trying to list on a public exchange don’t stop to debate it; they just use the terminology to entice investors with the sizzle, not concerned if there really is a steak on the plate. Academicians are more likely to dig into the discussion to try to actually answer the question with facts and logic. The first challenge, of course, is defining “intelligence”. Any such attempt must also define “consciousness” and further, identify the physical properties that form the foundation of that concept.
Alan Turing addressed this quandary more than eighty years ago. Turing was a mathematician and computer scientist who is known for decrypting Germany’s communication code machine, called Enigma, during World War II. In 1950 Turing offered a simple test to determine whether or not a computing tool could be said to represent “artificial intelligence”. Turing ignored the philosophical and scientific messiness of trying to define intelligence. He skipped past the challenge of explaining what consciousness is. Turing simply reasoned that if we can all agree that humans exhibit intelligent behavior, we need only compare the machine to a human. Turing said that, through conversational interaction, if you can’t distinguish the experience with a machine from that with a human, then the machine is indeed intelligent, albeit artificially created rather than grown through organic processes.
The Turing Test is now being applied daily by hundreds of millions of people around the world as they interact with these large language models. The results of employing such tools are already changing how intellectual work is being done. We may not know exactly when Turing’s test is satisfied. It could come as an accumulation of our interactive experiences over time. In the meantime, human-generated content (words, programs, solutions, products, designs, art, literature, concepts, etc.) is being joined, and often replaced, by non-human-generated content.
In past cycles of significant technology advancement that seemed to threaten the value of human economic activity (work), the displaced human workers transitioned to other work that technology couldn’t do better, faster and cheaper. This is why 70 percent of our country’s economic output is created by services rather than by product consumption. Manufacturing and agriculture became largely automated. Eighty percent of American workers now earn their living in the service sector. And human-created products and services have historically carried a premium of value above that of mass-produced offerings. Even when the result may be less perfect than what a machine could do. I know I prefer that my massage is delivered by an experienced and skilled person, not from time rented in a massage chair. I prefer music made by people with their bodies, not synthesized by a computer program, even if the musicians make a mistake occasionally. I prefer hand-crafted furniture rather than mass-manufactured products, if I can afford it.
I think this premium on human creation will continue to be the case as long as humans are still in control of their personal buying decisions. There may be a time when even that changes, of course. I dearly hope not. But if my conclusion is correct, it will be necessary to guarantee that humans were indeed the originators. I offer my official certification that any product or service provider can employ to assure customers that what they are buying has been created or will be performed solely by a human being. I am trademarking the logo you see above. To employ it, you must agree to the following: whatever product, service or creation you apply it to must have been entirely composed through the application of your brain and your body, or the brains and bodies of others. You can use tools. But you can’t compose a problem or set of design parameters and then submit those queries to a machine to manifest a result. You, and/or other humans, must provide the answers to the queries you design.
I’m not a Luddite. I enjoy and use technology daily. I also get frustrated with it and how it can often get in the way, rather than assisting our endeavors. It’s just that human-generated intellectual property is more valuable to me, and I think it may be similar for many other people. Now, with this certification, we have a way of assuring our customers, our friends, our audiences, our neighbors and the public at large that what they are experiencing is authentically and organically human created.
So if you wish to apply for the “AIN’T AI” certification, contact me at stan@stansewitch.com to receive the certification application. Send it back to me, signed by you and a witness with the required accompanying identification information, including a description of your creation. You will not need to disclose anything that is confidential and/or proprietary. I will review your submission, counter-sign if I’m convinced and send it back with permission to use the logo for the specific work. I will maintain the certifications so that anyone can inquire as to whether or not the applied logo was properly approved.
I wonder who will apply…?