* Disclaimer: This article was written by a human being, not an AI. As of today, we now live in an age when this disclaimer is necessary.
Part 1 of this series is here. It deals with the downsides.
Today I want to talk about the upsides.
Continued below the fold…
1. Professional Services for All
A whole variety of services that are typically only available to clients with money are about to become freely or cheaply available to anyone who wants them.
- Proofreading
- Language translation
- Graphic design
- Legal advice
- Investment guidance
- Tutoring
- Job coaching
- Therapy
- Interior design
- Image consulting
- Software development
- Etc…
What these services have in common is that a human expert with a storehouse of knowledge and/or a set of skills examines input that you provide in the form of text or images and then responds with some kind of output in the form of text or images, like factual answers, step-by-step instructions, a drawing, advice, or a written report.
The AI is going to be helpful in virtually any domain with those inputs and outputs. And soon enough this will extend to audio and video as well as text and images.
These expert-reliant services are typically expensive and scarce, which contributes to the cycle of wealth inequality. The experts are not a doorway to knowledge, but a locked gate. Only lawyers can afford interior designers and only interior designers can afford lawyers. What happens when literally everybody can afford both?
What happens is the world gets better.
(At least, that is what happens after the part where we cope with massive upheaval in the foundations of our entire economy.)
At some point, someone is going to develop a GPT-connected app that allows you to take a photo of your living room and the AI will give you professional-level advice on how to make it look nicer in terms of paint colors, decor suggestions, furniture layout, etc.
And this isn’t going to happen just for interior design. It will happen across every domain.
As a result, our homes will look nicer. Our money will be better invested. Our fashion choices will be enhanced with the help of an outside eye. More people will be able to take advantage of legal instruments like wills and trusts.
LESS OPTIMISTIC SIDE NOTE: Of course, this begs the question, what is going to happen to all the lawyers, financial advisors, interior designers, graphic designers, and so on? For some professions, the AI will aid rather than replace human workers. It will massively increase their output. And there will always be customers who want to work with human beings rather than use an app on their phone. But, yeah, lots of jobs are going to go bye-bye. Optimists will point out that automation in the past may have displaced workers, but it also created new jobs, and everything works out okay in the end. But I wouldn’t be so sure. The past isn’t a reliable predictor anymore. We’re on new ground and there are no precedents. The truth is, nobody knows what’s going to happen.
2. Medical Advancement
Robo-doctors
Let’s look at the S.O.A.P. model of medicine. It stands for Subjective, Objective, Assessment, and Plan.
- Subjective – the patient reports (or INPUTS) symptoms
- Objective – tests are run
- Assessment – the tests are reviewed to try to determine a diagnosis
- Plan – a treatment plan is given back (or OUTPUT) to the patient
What’s clear is that each step receives input from the previous step and generates output for the next step. All of this can be automated with existing technology — the AI just needs to be trained.
People will be reluctant at first to trust machines with their health. Human doctors should remain involved initially to provide oversight. But human practitioners are limited by their personal education, experience, knowledge, and powers of recall, where AI can be trained with vastly more data. It’s extremely likely that AI will not only make it possible to reach more patients, it will improve outcomes.
Medical Research
Medical research is labor intensive. As long as human hands are involved in the process, it will be slower and more expensive. We don’t yet appear to be at the point where AI can automate what bench-level pharmaceutical scientists do in the lab. But one aspect of medical research involves analyzing large amounts of data, and that’s something AI does well and has already been employed to do. A number of drug companies have been using AI for pharmaceutical research for several years.
Cancer cells escape detection from the immune system by producing adenosine, which is a common organic compound and one of the molecular building blocks of RNA. Adenosine is involved in various functions in the body. (Caffeine actually works to make us feel more alert by binding to adenosine receptors.) When adenosine binds to receptors on T cells, it prevents them from doing their job, thereby suppressing the immune system. One promising area of research in curing cancer is finding a way to shield the T cells from that adenosine.
An UK company called Exscientia teamed up with a German company called Evotec to look into that. Nature.com reports:
Where it might have taken the traditional discovery process 4–5 years to come up with the drug candidate—an A2 receptor antagonist designed to help T cells fight solid tumors—it was found in 8 months by harnessing Exscientia’s ‘Centaur Chemist’ AI design platform. This system can computationally sort through and compare various properties of millions of potential small molecules, looking for 10 or 20 to synthesize, test and optimize in lab experiments before selecting the eventual drug candidate for clinical trials.
Neil Savage, nature.com
Artificial Intelligence cut the time it took to develop this by 75%. The two companies announced last April that the new drug had entered human clinical trials.
In 2020, Google’s AlphaFold project used AI to solve the protein folding problem, The AI “can accurately predict 3D models of protein structures and is accelerating research in nearly every field of biology.”
A study from July 2021 already found “270 companies working in the AI-driven drug discovery industry.” As AI models improve, it follows that their impact on medical research will be greater.
Where might this lead? Cures for cancer, diabetes, heart disease, Altheimer’s? Almost anything imaginable will be attainable.
Aging
Cancer is a disease of the old. This graph of age-specific cancer death rates in 2020 from the CDC drives the point home:
Cancer isn’t the only disease of the old, of course. In this Moonshot Conversation on reversing aging, Dr. David Sinclair, professor of genetics at Harvard Medical School, argues that aging is the root cause of most diseases. We should think of these diseases as manifestations of aging.
In fact, there’s a debate about whether aging itself should be considered a disease.
So what does this have to do with AI? AI is already helping to discover anti-ageing chemical compounds. A 2019 article in Ageing Research Reviews about AI concluded that “AI technologies are rapidly emerging and are starting to deliver promising results in different fields of aging and longevity research.” For example, AI could help “identify key regulators involved in the onset of immunosenescence [the tendency for the immune system to deteriorate with age] and reveal the complexity of the interplay with other key biological processes. These regulators could, in turn, become targets for developing appropriate treatment.”
AI-Guided Regenerative Medicine
Induced Pluripotent Stem Cells can change into any type of cell through a process called cell differentiation. They can turn into blood cells, neurons, renal cells, or whatever you need. This potentially allows for the growing of tissue and even entire organs in a lab for transplant into the body. No waiting for a donor. And because the stem cells come from you, there’s no risk of the body rejecting the new cells.
But cell differentiation is complex, and there are factors involved that vary from patient to patient. Researchers are working on using AI to handle the volume of data needed for this technology to be possible at the clinic level.
3. Energy & the Environment
This week at the Lawrence Livermore National Laboratory, the US Department of Energy announced a breakthrough in nuclear fusion. Fusion is the process of crushing two hydrogen atoms together to create one helium atom, releasing a massive amount of energy. Fusion is what the sun does up there in the sky all day. Like nuclear fission, which is the process of splitting atoms (which is utilized by all currently existing nuclear power plants), fusion doesn’t produce any carbon emissions. But unlike nuclear fission, fusion doesn’t generate waste that remains radioactive for millions of years. And you don’t have to worry about Fukushima or Chernobyl-type melt-downs. The promise of fusion is safe, abundant, clean energy.
The DOE’s achievement was scientific energy breakeven. The energy created by the fusion reaction was greater than the energy of the lasers used to create it. There are caveats here, and naysayers bristle at the hype. The laboratory, pictured above, is the size of three football fields, and it’s a massive energy hog. The lasers delivered 2.05 megajoules of energy to their target, and the ensuing fusion reaction generated 3.15 megajoules, which is great. But this didn’t come anywhere close to producing the 300 megajoules needed power up those lasers. There are still many hurdles to overcome to commercialize fusion energy. Nevertheless, this is a turning point in the fusion story.
One part of that story is AI, which is already helping to accelerate fusion research. The experiments conducted at LLNL are expensive, so LLNL uses neural networks on special computers to run simulations. A week before the successful experiment that made the news, CogSim, LLNL’s machine-learning model, predicted the experimental design would work.
Another approach to fusion uses magnetic fields to control plasma in a donut-shaped contraption called a Tokamak. Google’s AI is helping with that. (See here and here.)
And if that isn’t good enough for you, AI shows promise in a vast array of other ways to help fight climate change: How To Fight Climate Change Using AI (forbes.com)
In Conclusion
To the three great technological revolutions–the agricultural, the industrial, and the computational–we will add a fourth: the AI revolution.
Sam Altman
The CEO of OpenAI, Sam Altman, wrote an article called Moore’s Law for Everything.
SIDE NOTE: I’m glad there are people like Sam Altman stewarding the development of AI, because he seems genuinely interested in being ethical and getting AI right. That said, I find his optimism a bit unnerving. One reason I like Joe Biden is that I have an easier time trusting a weathered leader, someone who has failed, been beaten down, made mistakes, and had some of the idealism knocked out of her or him by life. Reality is complex, and if you aren’t a little broken by that complexity then I don’t feel like you’re here with me. At moments Altman’s bright-eyed utopianism seems dangerously naive.
Anyway, Altman is right that we’re on the precipice of a technological revolution, and he says several smart things about it in that article. Here’s the spliced-together crux of his argument:
Consider the example of semiconductors and Moore’s Law: for decades, chips became twice as powerful for the same price about every two years. AI will lower the cost of goods and services, because labor is the driving cost at many levels of the supply chain. Imagine a world where, for decades, everything–housing, education, food, clothing, etc.–became half as expensive every two years.
I haven’t yet mentioned other fields where AI will be helpful, like engineering, education, product design, and space exploration. It could have a big impact in surprising areas, like matchmaking. It will have upsides we just can’t predict at this point.
What AI promises is nothing less than rapid advancement in virtually every aspect of life toward a Star Trek-like future, where people live together more harmoniously in an environment that is healthier, happier, safer, cleaner, more abundant, and more beautiful, and where technology makes it possible for everyone to enjoy lives full of meaning and dignity.