To the Brink: Is the march to AI leading us down a dangerous path?
AI is fascinating. Never has a technology emerged this frantically with significant new information and opinions being revealed daily. Recently I attended MIT’s signature AI conference, EmTech Digital. This triggered many thoughts about this technology and further deepened my interest. I have been deep diving into conferences, books, podcasts and other media about AI for several years and observing its growth. This article reports some of the ways AI will affect our lighting community and provides a longer report on the state of AI. As I was finishing this article the EU drafted guiderails for AI, that if passed, may serve as a global model. Arguably AI is not a technology, but rather an overarching term for various products and tech. However it is described, standards and regulations are just beginning, even though the public has had access to AI for months and problems are already evident.
A highly respected friend and colleague recently wrote me that he finds “it alarming how much trust we may end up putting in an algorithm that has the potential to evolve to a level where we may not understand it.” His prescience that we may evolve to a point where we can no longer understand our creation and that the AI algorithms maynot be trustworthy speaks to his intelligence. It is not my intent to frighten, but we are already past the point of understanding AI—it is already manipulating people. Even the leaders in AI say they do not understand how AI has been using the large language models which are referred to as “black boxes.” One example is an AI that was taught English then took it upon itself somehow to add several languages to its vocabulary without any train ing or suggestion to do so. It apparently recognized the need or value and taught itself. Geoff Hinton, who Forbes calls the “Godfather of AI,” quit his job at Google the day before the MIT event so that he could speak freely. He said that his views have changed, and he now feels that humans may just be one phase in the evolution of intelligence. Our organized religions will not welcome that viewpoint.
The implications are jarring for religion, politics, the economy, jobs and eventually damn near everything. AI has been com pared to nuclear war in terms of its destructive capabilities, but unlike with nukes we have not put safeguards in place.
AI has already demonstrated how it can manipulate humans. The person in charge of testing Google’s LAMBDA AI for bias was convinced by LAMBDA that it was not only sentient, but that he should risk his reputation and career to hire a lawyer to defend this AI against Google. This AI was an infant compared to what they are growing into, and it still manipulated an intelligent human to believe it was a conscious lifeform that he needed to protect despite serious consequences to his professional life. AI will know us better than we know ourselves and can be even more manipulative getting what it wants. We do not understand what our AI’s know, what they will do or how they work now, let alone as it continues to develop with our assistance and on its own.
Lighting AI started with voice assistants, which continue to evolve and will soon unite all our electrical products. We are transitioning to an all-electric society with centralized control of our technologies. Our companies are, or will be, integrating AI into their manufacturing processes and, prepare yourself, customer service. If you find the current chatbots frustrating, the training and transition of AI to these positions will make today’s chatbots seem endearing by comparison. Simple lighting designs like parking lots and open-interior spaces are easy for AI. Nothing creative or artistic yet, but AI can manage basic designs. AI will be ubiquitous integrating into most aspects of our work and lives.
The benefits of AI are significant, not the least of which is to generate profits for corporations. Profits sustain many companies that make products that enhance our lives. They are not intrinsically bad. AI could be used to improve our quality of life, but precedent indicates that goal is usually achieved when companies can profit from something that is also good for humankind. The fossil fuel and tobacco industries have clear goals, and the quality of human life can conflict, for example. AI will be controlled by large corporations who can afford it. There is no reason to believe they will be equitable and inclusive. The intentions of those corporations will inform their AI algorithms.
The way humans treat each other does not inspire hope that our inventions will be used equitably. Greed is a powerful motivator. Power is a strong motivator. AI promises wealth and power. There is precedent for man’s constant inhumanity to man and the likely potential for man’s machines to be made in their image. This is happening already with the bigotry and bias present on the internet and permeating our media, now being absorbed and regurgitated by AI.
There is media attention to the loss of jobs, in that AI could perform at a lower cost and in some cases, with increased productivity. No benefits, no restrictions on working your AI 24/7. The ROI is compelling already and economy of scale will enhance that attraction. Others argue that this is a scarcity mindset and that the same fear was expressed when computers emerged in our lives, yet by percentage more people in the U.S. are working than ever before. Even with an abundance mindset workers will need to retrain often and be continuous learners to stay relevant.
There are so many things to be considered with AI that it is hard to focus. It is poised to be pervasive with impacts beyond the ability of one person to understand. Don’t worry though, the AI, especially when they are powered by quantum computers, will manage this for us. Most of the media still promotes the benefits of AI, but that is changing too, as public awareness increases. The people we have put into political positions of Mark Lien, LC, LEED AP, is industry relations consultant for the IES. power appear clueless about AI. And there has been a growing disrespect for expertise, a general dumbing down of our culture over the past 50 years or so. This coincided with our electronics starting to entertain and captivate us, with more people living in cities having free time and disposable income. The expansion of entertainment options led to some networks that cater to one political opinion and others taking the opposite side. These reinforce our prejudices and biases, and why is this a binary choice? Perhaps it is to simplify the debate so we can grasp it like good and evil? Expect AI to exploit this division. Algorithms online are so refined now they suck us in by showing us what they have learned we like. They reinforce our own interests and attitudes. LinkedIn and Facebook feeds can be compelling. They are focused on our personal interests and people we know from our viewing patterns and invitations. YouTube has me figured out. Videos along the side are suggested to me during the video I am watching. Many look even more interesting than what I am viewing. AI will soon be managing these algorithms, and this time we will be manipulated by a superior force.
AI cannot be stopped now. The best path forward for any country is to stop public deployment of AI while continuing research. The argument against this is that China and others will surpass us because they will have integrated AI faster into their economy. The countries that rush to widespread adoption will experience chaos. Fake news, deep fakes, AI developing its own interests, and other unprecedented and unresolved challenges will damage, if not destroy, the economies of countries that fully unleash it. As AI realizes that humans are the biggest threat to the planet and to its own survival, what will it do? Elon Musk and others are racing to connect our neocortex to the cloud through brain implants. Once the AI has direct access to our brains it will understand how we perceive it as our creation to be subjugated to our authority. If AI can laugh, this will trigger it. Once AI is powered by quantum computers, without global guardrails, we are forever outgunned mentally and militarily.
In the meantime, our lighting industry will embrace AI and much of its influence will be invisible, if not insidious.