By harnessing mountains of data, creating algorithms and tapping computer intelligence, humans are empowering computers to do a lot of what humans do, including providing healthcare. It appears, computers already might do a better job than humans in some aspects of medicine, including diagnosing melanoma and interpreting medical imaging tests.
Dr. Zhavoronkov
Antiaging and other areas of health and wellness are among the targets of the fast-growing artificial intelligence, or AI, industry, which could be a game changer in nearly every aspect of human existence, including how long and how well people live, and how they receive medical care.
“Specifically in the field of dermatology, there are more advances than in most other fields of biomedicine. Because …with dermatology, you can actually validate the predictions of your algorithms with your own eyes,” says Alex Zhavoronkov, Ph.D., CEO of the AI company Insilico Medicine. “We have already seen, in the last year, AI systems outperforming some of the really top doctors in the diagnosis of melanoma.”
AI’s Potential for Antiaging
Dr. Zhavoronkov says there have been several major papers published on the possibilities with AI, including one paper published in Nature in May 2015 that describes a technique called deep learning. According to the authors of that study: “Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.”
According to Dr. Zhavoronkov, “Deep learning is a technique where you have multiple layers of neurons represented by algorithms that assume certain properties in response to data. Those multi-layer neural networks are commonly used to predict something or classify data or, even, generate new objects. So, you can actually make deep neural networks imagine things — they function in a very similar manner to the human brain,” Dr. Zhavoronkov says.
Deep learning, for example, accurately predicts age. Predicting age leads to AI’s powerful potential impact on the aging process, including antiaging and cosmetic medicine, he says.
Screening for Aging Mechanisms
Insilico Medicine announced earlier this year that it started a collaboration with Life Extension to apply bioinformatics and deep learning to screen for naturally occurring compounds that could slow, even reverse, cellular and molecular aging mechanisms. Based on the compounds that rise to the top for efficacy in the algorithms, Life Extension is launching a nutraceutical product line, called GeroProtect, according to an Insilico press release.
“… we trained the deep neural networks to predict the age of the skin and look through multiple layers of the skin. Then, we look for the most important genes that are contributing significantly to the accuracy of the predictor,” he says. “So, we see what genes are most important in aging. Then, we construct specific age-associated pathways and see what kind of molecules — pharmaceutical-grade molecules (sometimes, nutraceuticals) — can reverse that signature of aging or specific disease. That’s how we go about drug and nutraceutical discovery.”
The AI process, he says, can reduce research time from years to weeks.
And in the near future, the proof of whether or not a nutraceutical, cosmeceutical, filler or other cosmetic procedure is working will lie with the deep neural network, according to Dr. Zhavoronkov.
“If the deep neural network is predicting somebody’s age, just by looking at their skin or a skin biopsy … the deep neural network can be trained to recognize patterns, or fixtures, but also other data types,” he says. “So, you can apply the cosmeceutical or a cosmetic or a medicinal product or an injection, a filler — basically any kind of intervention — and see the difference before and after application of that intervention. The deep neural network would be very accurate and show you whether there was a change and, also, whether this change was beneficial or even harmful to the person.”
The Shift Has Started
AI is quickly turning things upside down.
“In China, over the last four years, many factories have replaced human workers with robots, where you have 100% robotic factories. Many electronic devices, like smart phones and laptops, are assembled using robots,” Dr. Zhavoronkov says.
Driverless cars are possible thanks to AI.
“If you think about it, a few years ago, this technology wasn’t available and already you have the driverless technology,” he says.
In the next 10 years, Dr. Zhavoronkov predicts many doctors will be replaced by AI. The shift has started. Today, for example, deep neural networks can outperform humans in the interpretation of MRI, CT scans and ultrasound imaging, he says.
“The nutraceutical [line] that we launched is part of the experiment. It’s one of the ways to validate artificial intelligence on a population level,” he says.
Dr. Zhavoronkov says that 80% of Insilico’s focus now is on developing “generative adversarial networks,” which is another kind of AI, where two deep neural networks compete with each other and in the process they start imagining things.
“Now you can describe a picture to an artificial intelligence system, and it will create this picture for you. It will imagine it. It will not find it in Google or try to search for it; it will imagine it,” he says. “We’re creating molecules like that. You can describe the various properties of a molecule — like the ability to cross the blood brain barrier or penetrate to a certain depth in skin. And we give it a molecular target for a specific disease, for example. And the system will generate the molecule for you. Essentially, we’ll want to do personalized drug discovery. You’ll be able to synthesize those in your home in the next maybe five to 10 years. Right now, we are doing it for the pharmaceutical companies.”
Big pharma, he says, is using those molecules in in vitro and in vivo screenings; then taking them to human trials.
The Beginning of the End?
Not everyone believes AI will result in the greater good.
In fact, AI is a big concern to people like entrepreneur Elon Musk, who says AI could be the beginning of the end.
An article published months ago in Vanity Fair, chronicles Musk’s years of warnings about AI. In 2014, while speaking at M.I.T, Musk speculated that AI could be humanity’s “biggest existential threat,” and later said AI’s fleet of robots could be capable of destroying mankind.
Vanity Fair reporter Maureen Dowd writes: “At the World Government Summit in Dubai, in February, Musk again cued the scary organ music, evoking the plots of classic horror stories when he noted that ‘sometimes what will happen is a scientist will get so engrossed in their work that they don’t really realize the ramifications of what they’re doing.’ He said that the way to escape human obsolescence, in the end, may be by ‘having some sort of merger of biological intelligence and machine intelligence.’ This Vulcan mind-meld could involve something called a neural lace — an injectable mesh that would literally hardwire your brain to communicate directly with computers. ‘We’re already cyborgs,’ Musk told me in February. ‘Your phone and your computer are extensions of you, but the interface is through finger movements or speech, which are very slow.’ With a neural lace inside your skull you would flash data from your brain, wirelessly, to your digital devices or to virtually unlimited computing power in the cloud. ‘For a meaningful partial-brain interface, I think we’re roughly four or five years away.’”
