Of the many great paintings by Michelangelo on the ceiling of the Sistine Chapel, one stands above all others as his most recognizable work. “The Creation of Adam” depicts God with hand outstretched to his creation, Adam. The painting raises a big question in the world of AI and Big Data—can humanity safely outstretch our collective hand to AI-controlled robots and machines, or are we likely to get our arms ripped off by imprecise electronics?
Today most people still think of AI as a mere computer program, solving problems in the digital realm. A formidable application of computer science, it can assist us with once inconceivable tasks, such as detecting cancer and analyzing data far too complex to wrap our minds around. Those AI applications, while still critical, no longer represent the state of the art. The cutting-edge work in AI now centers on how to safely introduce devices controlled by machine learning into our physical reality.
AI already has begun interfacing with the physical world in many ways, mostly in the industrial and farming sectors. For instance, GAIA (Geospatial Artificial Intelligence for Agriculture) is a system implemented in Australia combining imagery of vineyards and AI to inform farmers on the health of their plants on a row-by-row basis. Yes, it can even predict the quality of their fruit! And as I wrote for Forbes last year, high-tech manufacturers, such as Aerion Supersonic, are using “digital twin” tech to create an AI-driven mirror of their supersonic jets and manufacturing plant, tracking real-world performance down to the minutest detail.
But these examples concern large scale industrial applications. The more critical challenge facing us is this: Will AI ever be safe for the humans? It appears so. Companies already are working on the question of AI for personal use in the physical world. WineCab developed the first industrial-strength robotic arm for the home that can delicately handle its owner’s wine collection. (The system is safe enough not only to negotiate precious bottles, but also to work in close contact with humans reloading its storage compartment.)
While WineCab is busy mastering this specific application of AI-controlled robotics, industry leaders must solve two problems before mastering the domestic market: accuracy and precision. Often used interchangeably, these terms describe different factors vital to creating a world in which AI makes a positive impact on our lives.
Let’s break them down individually. Accuracy concerns how close to the target a value is (i.e., the bull’s-eye on the target), while precision measures how close repeated actions are to the original (i.e., tight grouping in target shooting). In the 21st century, we have come to rely on technology to be both accurate and precise. After all, what good would GPS be if it wasn’t accurate enough to lead us to our destination? Similarly, it would be maddening if our GPS determined the right destination yet chose imprecise routes to get us there.
As accurate and precise as GPS is even now, the future of AI requires far greater precision. For instance, GPS may be great at helping us get from place to place, but it’s still too inaccurate and imprecise to drive our cars for us. (Safe driving is a matter of inches, and GPS is accurate to roughly 16 feet.)
It’s therefore little wonder that top players in the field utilize far more accurate and precise sensors, such as Argo AI, employing Lidar sensors in its partnership with Ford for self-driving cars. Lidar and other technologies may seem perfect for tomorrow’s vehicles but are unlikely to be useful for AI applications pertaining to the human body (at least at present). The question of how to bring greater accuracy and precision to this type of real-world application illustrates a deeper problem: If we aren’t careful, we’ll hit a wall on the usefulness of this technology for humanity.
The dilemma facing AI is easy to understand. By its nature, AI’s applications are infinite—in the digital realm. As long as we can increase our storage capacity, datasets can grow exponentially. But the world we live in is the opposite—its physical resources of all kinds are finite. To bridge technology from the digital to the practical, devices must become ever more precise, or we won’t capitalize on AI’s potential to improve fields with significant human impact, such as medical care.
We are relying on the engineers of today to design hardware to support AI digital precision for the future. But what emerging tools are innovating for this technology bridge? One company is harnessing the power of something called “piezoelectricity” to create the level of accuracy and precision our future requires. Unfamiliar with this term? You’re not alone. And yet piezoelectricity is not new science. (The Curies explored the piezoelectric properties of crystals way back in the 1880s.)
“Piezo” comes from the Greek word “piezein,” meaning “to squeeze.” The piezo effect occurs when force is applied to certain materials, generating an electrical charge. You may be thinking. “Okay. That’s nice, but what does this have to do with me?” Well, with Memorial Day nearly upon us, backyard grill masters from coast to coast may be interested to know that each time they fire up their grills with a push button starter, they’re activating a piezo element. Similarly, most folk musicians amplify their music with a piezo pickup on their acoustic guitars.
AI applications won’t use this effect to precisely control the real world. Instead, they will rely on the reverse piezo effect. When the piezo effect is reversed, electricity is added to ceramic piezo to create motion. One company, Piezo Motion, has emerged as the industry leader in building ultraprecise piezoelectric technology at scale, helping to revolutionize how AI interacts with the real world.
Already, the difference in precision between a standard electric motor and Piezo Motion’s piezoelectric motor is startling. Consider the task of forming a circle. A high-precision electric motor will break this circle into 30,000 tiny movements. Not bad, right? Well, a piezoelectric motor can isolate the same circle into 600,000 tiny movements, providing an AI system with an astounding level of precision, not to mention a largesse of performance data to analyze.
What’s more, Piezo Motion has developed piezoelectric motors capable of controlling movement—down to the nanometer. “That level of precision becomes quite important when we consider what AI holds for the future of medicine,” explains Hassan Kotob, chairman and CEO of the company. “As AI begins to play a larger role in assisting doctors with microscopic robotic surgery, ultraprecise control over repetitive motions will be required, necessitating piezoelectric motors with incredible new levels of control.”
Building on this promise, ultrafine precision isn’t the only benefit piezoelectric devices portend for the medical field. Piezoelectric motors can be manufactured from completely nonmagnetic materials, a key factor for imaging technology, such as MRI machines involved in medical treatment. Also, the motors can operate silently and sans energy consumption when not in motion.
But the promise of piezoelectricity upon AI doesn’t end with motors. Already, Piezo Motion is manufacturing its technology into healthcare applications such as micro-dosing pumps that may soon disrupt the pharmaceutical industry. “Imagine for a moment an AI that knows precisely what dosage of medication a sick patient should receive, down to the milliliter and is able to dispense it accurately and safely,” says Kotob.
So, can we precisely measure that level of dosage? With a piezoelectric pump, the answer is quickly moving toward “yes we can.” Replacing the traditional pump drivetrain with piezo components enables the precision dosages that an AI system demands. Further, the company’s vision of pharmaceutical delivery is a 3D-printed, wearable device administering medicine in exact dosages at times specified by an AI overseeing treatment. This makes sense. Humans are not “one size fits all,” so why are medicines still being delivered in generic dosages across patients?
The synergy between piezoelectric motion and AI systems involves other matters of life and death. As the quest for self-driving cars ramps up, we can imagine the necessity for Herculean feats of satellite precision and triangulation to stop one’s vehicle should something suddenly appear on the road. Extrapolating further, we can envision a day when smart farming equipment employs breathtaking accuracy to pinpoint the exact amount of pesticide to spray on insects—without harming crops or wildlife. It also doesn’t hurt that piezoelectric motors offer tremendous energy and savings over conventional motors, and the low cost of the latest generation of Piezo Motion motors makes them a fit for practically any sort of project.
Ultimately, as the next generation of engineers and computer scientists discover novel applications of AI to better conduct our lives, one thing is clear: Our physical tech must keep up by achieving increased accuracy and precision demands. Piezo Motion’s piezoelectric motors and pumps offer one more arrow in the quiver toward realizing a better future through human-centered innovation. What bright vistas await us in the years to come? This is uncertain. What we can suggest is that with once unimaginable resources in our hands, a more astonishing world is sharpening into focus. And faster than we think.