The magic of AI rests in its uncanny ability to optimize a given goal in a specific way for each person. AI trains “neural networks” on huge amounts of labeled data; then they use what they’ve “learned” to mathematically pick out and recognize incredibly subtle patterns within other mountains of data, to maximize the given goal. What if we could use the magic of AI to make us healthy or happy? This is quite doable in the next twenty years, and one of the key points I explore in my new book, AI 2041.
How can AI make us healthy?
AI can work with scientists to invent new drugs that can find treatments for rare diseases. For example, biotechnology company Insilico Medicine has produced a tool that helped scientists find a cure for pulmonary fibrosis much faster and cheaper than before. AI works best when we can take voluminous data, measure, optimize, and improve. And more personal healthcare data is being collected than ever – from wearable devices, blood tests, cancer markers, MRI scans and genetic sequencing. AI can make use of all this data in many ways. AI diagnosis can learn from billions of actual cases, treatments, and outcomes, and find effective personalized treatment for each patient. The ability to personalize, and to create the perfect memory of all new drugs, treatments, and studies, will cause AI to exceed human doctors before 2041. I also expect that AI will be able to look at health data globally and find dangerous potential pandemic outbreaks – something on everybody’s mind today.
A detailed health examination including radiology, blood test, and genetic sequencing can be fed into AI to compare against billions of other cases. I use an AI app called young.ai, which, along with an expert doctor, can interpret data, and recommend personalized changes in lifestyle, sleep, food, nutrients, and medicines to keep each person healthy. For example, they (AI and my doctor) diagnosed that my sleep quality was low due to dependency on Ambien, and then they helped me find a solution to fix this problem at the root. Also, they replaced my vitamins with a new set of nutrients including geroprotectors, with positive results. After using this service for one year, their advice has already made my blood appear “six years younger”. Human life expectancy has increased about 15% every twenty years for the past century. With the help of AI, we can expect this trend to continue, resulting in our longevity and improved health and quality of life in 2041.
How can AI make us happy?
Can this approach be extended to make us happy? One of the stories in the book AI 2041 is about a Middle Eastern monarch who wants to use AI as the elixir for happiness. The monarch summons an eclectic group of guests to explore this fascinating problem …on themselves. Each guest voluntarily agrees to contribute all personal data, trading privacy for happiness, in an experiment to build a personal AI that makes each guest happy.
For this experiment to work in real life, we would have to collect much personal data to help us measure happiness. Our wearable devices know our heartbeats per minute and our blood pressure. Our phone or computer can see our face and see our smile, facial expression, and micro-expressions (usually undetectable even by people). Our emotions can also be measured by the hue of different parts of our faces, which is caused by localized blood flow, and the pitch, loudness, tempo, emphasis, and stability of our voice. In addition, the trembling of our hands, dilation of our pupils, the welling up of tears, patterns of blinking, humidity of skin (pre-sweating), and changes in body temperature are all useful features to estimate someone’s state of mind.
In addition, hormone levels can be used as another approximate measure of happiness. Serotonin is correlated with well-being and confidence, dopamine with pleasure and motivation, oxytocin with love and trust, endorphins with bliss and relaxation, and adrenaline with energy. Soon we will be able to make a transdermal biosensor membrane with a matrix of under-the-skin microneedles and an electrochemical sensor that continuously measures these hormones. In the longer term, we will be able to measure the electrical activity and the architectural component of our brain. That is what BCI (brain-computer interface) companies like Elon Musk’s Neuralink are trying to do, though these may take longer than twenty years to become commercialized and legally approved for use.
Back in the story, the monarch builds an island with stimuli (e.g., a VR adventure, a soothing concert or a friend introduction) that will maximize each inhabitant’s “happiness” measures. For example, someone who wants to make friends with similar interests is matched to others with the same desire. Someone who wants an adventure is put into life-like VR adventures. Everything described in the story will work in 20 years. To the monarch’s chagrin, the island brings pleasure to people, but the experiment does not create a utopia, because these metrics are only an approximation of short-term happiness, not long-lasting.
But what is happiness, really?
The monarch’s high-tech experiment falls into the trap of the “hedonic treadmill”. Psychologist Michael Eysenck introduced the term “hedonic treadmill” to describe our tendency to always readjust to a fixed level of happiness, despite monetary and possession gains (or losses). Studies have shown that people who come into sizable wealth (such as winning the lottery) are happy for a few months, but after that, their happiness usually drops down to the baseline level before they came into wealth. In the AI 2041 story, the island guests initially indulge in various pleasurable activities that produce short-term bursts of happy feelings, but over time, they are back on the hedonic treadmill, always treading, but never achieving lasting happiness.
So the monarch learns along with the readers, that to reach lasting happiness, we need to move up the Maslow hierarchy, and attain the sense of belonging, love, esteem, and self-actualization. These higher levels of the Maslow hierarchy don’t involve moments of instant gratification, but rather the long-term pursuit of meaning and purpose.
Can AI help us find love, meaning, and purpose? This is unlikely in twenty years, as we do not have any grasp as to what these mean in terms of our physiological manifestations. What we do know is that in 20 years, AI will be able to measure and improve human health as well as help us have more happy moments. Knowing that, we can work to make AI subordinate to human needs rather than human greed. And that in itself will be monumental progress. Kai-Fu Lee is a co-author of AI 2041: Ten Visions For Our Future from which this article is adapted.