Artificial intelligence (AI) burst into the public discussion in the early part of 2023. Along with the overly optimistic views of what it would do for humankind came overly pessimistic predictions of dire consequences. As is the usual case, reality lies in the middle. What should citizens understand about this technology, and what should they do?
What is artificial intelligence of the kind in the news today? It is another form of information technology. It is not some magical, almost-human facility, as some pundits are suggesting. At the risk of oversimplifying, the large language models (LLM) that are a core element of ChatGPT and similar systems could be viewed as an extension of big-data methods which have been expanding for years.
These AI systems are formed when the knowledge engine, the actual software, processes enormous amounts of information located on the web and elsewhere. It is important to know that the “smart” engine would be useless without this vast supply of information from which it “learns.” (I have used quotation marks to emphasize that the meaning of these words as applied to AI are different from how we use them applied to humans.) As an interesting side note, these AI systems require tremendous computing capability which requires a great deal of energy, so much so that some companies involved in the development of AI systems have invested in nuclear energy ventures.
For the typical citizen acting in their role as citizen-ruler, the above is all the technical knowledge that is needed to cast the necessary political, economic, and social votes that will be necessary to guide the use of AI systems.
The End of Civilization
Yes, that heading is a bit dramatic. However, multiple individuals and institutions issued that actual warning in 2023. When I first began reading these dire, existential warnings, I assumed that the people behind them were making the usual power plays. For example, they wanted to control the entrance of new competitors, influence politicians and regulatory bodies, and convince the public that only experts could be trusted to save the world.
Soon however, I understood that some of those who issued such warnings were serious. They believe that the AI systems will soon become not just sentient but volitional. As such, we cannot just “unplug the power” because the system will prevent us from doing so. There are too many things wrong with that line of thinking to address in this article. Suffice it to say that humans design and build AI systems as well as the energy sources that power them. Thus, AI systems have only the power we grant them.
Impact on Politics
Artificial intelligence will be used—indeed, is being used—to create realistic but false communications. In this way it poses an indirect threat in the political arena. Our first step is to acknowledge that the use of misleading or deceitful information is nothing new in politics. It has been part of the political process since our founding. Among those looking to advance an agenda or gain power, some will always use the latest technology to create and communicate misleading information. This was true for newspapers, radio, television, and social media.
Citizens always face the challenge of deciding what is real and what is fake. Imagine what it was like two hundred years ago trying to figure out whether a newspaper article was indeed written by the presented author and whether what it said was true or accurate. Prior to this current generation of AI, we already had tools to edit photographs and videos to present what was not accurate. Artificial intelligence adds two more capabilities to the challenge. First and most powerfully, it provides tools for a deceiver to create written material that matches the style of any person, audio that sounds like another person, and video that shows a person doing something they never did. Second, it allows a deceiver to create misinformation quickly.
What can citizens do about this? As it turns out, there is a quite simple thing that can render most of the deception irrelevant. That thing can be captured in one word: accountability. Some sources of information are easier to identify and hold accountable than others. If a major newspaper, magazine, or television station publishes something through their official channels, we can know with some certainty who is responsible. Of course, that does not mean the source will not try to deceive us, but when a source can be identified and when it owns assets and a reputation, citizens have more options to hold them accountable. A natural corollary to this solution is to ignore all other sources of information.
Artificial intelligence can be used by deceivers to deceive us only if we allow them to do so. If we are diligent in knowing with certainty who produced the information and if we are willing to hold them accountable politically, economically, and in the courts, we can neutralize the impact of AI on politics. On the other hand, if we are unwilling to invest that effort—if we are unwilling to be responsible citizen-rulers—we can expect to be deceived.
Impact on the Economy
As with any other new technology, AI will affect the economy. In the long-term, we hope that each new technology will ultimately benefit everyone. Even if that is the case, in the short-term, there will be winners and losers. How new technology comes about, its rate and nature, and how it is accelerated by our monetary and financial structure, is a broad and important social concern but beyond the scope of this article.
Instead, we can consider the unique aspects of new artificial intelligence technology. The most important economic effect is the fact that much of its value comes from processing information that belongs to someone else. AI systems of the type in the news today are useful only because they have been trained by processing information which was produced by other people at their expense. It could be compared to a manufacturing process in which the manufacturing company did not have to pay for the raw resources because the company could take them from someone else who already paid for them.
Some owners of information, that is, owners of intellectual property, have already filed suit to get compensation for the use of their information. The battle for ownership of intellectual property will significantly influence the way AI affects the economy.
As a simple example, suppose that AI engines learn from many resources and then turn around and provide answers to their paying customers. It could fairly be said that the AI system is monetizing the work of other people. If that is allowed to continue, current producers of information will stop doing so. Or the producers may develop technology to prevent AI systems from accessing their information.
Impact on Society
How will AI impact society beyond the political and economic realms already discussed? One way is already taking place. Some companies are employing AI technology to create fake representations of actual individuals. The systems work by gathering photographs and videos of the victim along with audio and textual information. From that information they can generate realistic representations that are completely fake.
Those in power, especially in government, will try to use AI to monitor and characterize citizens. They will present these actions as providing extra security, prosperity, and harmony. If citizens allow this to happen, they will soon see restricted freedom.
What might citizens do in response? The first thing is to treat all representations in the same way as discussed above for the political realm. That is, ignore all representations that are not certifiably authentic. The primary reason that companies generate fake representations is for attention or money. If no citizens paid any attention or any money to fake representations, they would diminish greatly.
Second, comparable to the intellectual property issue noted above, we must decide who owns what and then enforce those laws stringently. For example, as a society we may decide that each person completely owns all representations of themselves and all data related to themselves. Any use of such information without the explicit approval of the individual would be illegal and subject to stiff penalties.
Third, citizen-rulers should restrict the use of AI by those in government until the exact nature of the use is understood. Citizens should prohibit its use by agencies without explicit approval by Congress and the president. Any use by any area of government must be transparent and subject to clear, stringent enforcement and penalties. Citizens can insist on analogous protections from business.
How Citizen-Rulers Can Respond
A central theme in all my work on this website is the centrality of the citizen-ruler functioning with proper understanding. I emphasize applying fundamental principles with wisdom, coupled with competent and honest analyses. That is why as individual citizens, we do not need to understand any more technical details about AI technology than was given earlier in this article. Rather, we need to take responsibility for the information we consume, and we need to support strict enforcement of property rights. Those two actions alone will prevent or mitigate many of the negative effects that might come from the use of AI systems.
Implicit in those two actions is learning what information sources are clearly identifiable and can be held accountable. Stated differently, as citizen-rulers we must take an intense interest in helping develop and support trusted or primary sources of information, without which we are hindered in our role as citizen-rulers.
In closing, I want to be clear that the science and technology underlying AI systems can be of benefit to society. I focused on the potential problems for three reasons. First, that is primarily what has people’s attention right now. Second, we know from fundamental principles that those in power will use any new technology to increase power. Third, only if we address the legitimate problems attendant with uses of the technology can it deliver its potential value.