With technologies at such a small scale, we’ll have access to new places and new kinds of data. The rise of tiny machine learning or TinyML – the field of machine learning technologies capable of performing on-device analytics of sensor data at extremely low power – further expands the possibilities for data aggregation, management and personalization. These developments could revolutionize:
Healthcare. Recently, researchers used 3D micro-printing to develop the world’s smallest, flexible scope – less than half a millimeter across -- for looking inside blood vessels and providing high quality 3D images to help scientists better understand the causes of heart attack and heart disease progression.2 Cornell University researchers recently unveiled a laser-activated robot that is the size of a paramecium, can fit almost anywhere in the human body, and is inexpensive enough to produce at mass scale.3 The hope is that one day swarms of these robots will swim through bodily fluids clearing plaques, repairing blood vessels, and even probing into grey matter.4 Elon Musk’s brain-machine interface company Neuralink recently showcased a pig, Gertrude, outfitted with a device that recorded signals from an area of her brain linked to her snout. Musk is hoping that these devices could eventually solve numerous neurological problems, such as memory loss, strokes, and addiction.5
Environmental Science. Researchers recently built a tiny camera which is about as wide as a thumbnail to learn more about how insects react to their environment.6 Space startup Swarm recently got approval to begin operating its satellite connectivity service in the US.7 Their satellites are even smaller than most CubeSats, roughly the size of a sandwich at 11 x 11 x 2.8 cm. Their goal is to provide a worldwide, affordable satellite data network that will be suitable for use in IoT applications, including maritime and ground logistics tracking and agriculture.8 Warwick University has been exploring the benefits of using CubeSat’s tiny satellites to provide a cost effective means by which developing nations can monitor the herd size and migration patterns of endangered species.9
Product Design. Startup Wiliot is developing tiny Bluetooth stickers that measure temperature, location, and air pressure, and can transmit data back to the cloud which could help with inventory management and authentication of products.10 Putting these stickers on clothing could help determine which items are a hit with shoppers, even after they left the store.11 Since COVID-19, they have turned their attention to bringing item-level traceability and temperature sensing to individual vials of vaccine.12 The Neviano UV Protect swimsuit is equipped with a removable medallion-style waterproof sensor – about half the size of an adult thumb – that sends alerts to the wearer’s iOS or Android device when UV levels are high and more sunscreen should be applied.13
Public Safety. Google Pixel Buds, using TinyML, can already alert you to the sounds of a crying baby, a barking dog, or the siren of an emergency vehicle when you’re listening to something and may not hear the sound.14 Swim.AI uses TinyML to route traffic more efficiently.15 As California copes with one of its worst wildfire seasons, PG&E Corp. and other major utility companies are investigating how AI could reduce the risk of future conflagrations by using drones with AI capabilities to spot potential equipment issues such as worn or rusty hardware.16
Accessibility. Researchers have developed a brain implant for blind people that bypasses the eyes and allows rudimentary vision.17 To assist amputees, the AI-powered Avocado Wrist Connector can be fitted to any upper limb prosthesis to improve use.18 Researchers in Toronto have developed a tiny pill-sized “heater” that could allow resource-limited regions around the world to test for infectious diseases without the need for specialized training or costly lab equipment.19
The Dark Side: A New Arsenal for Surveillance?
These new, small and smart technologies are hitting the market just as widespread suspicion of corporations and government agencies is on the rise.20 We should expect increased concerns with respect to:
Privacy. Small cameras on insects may be a great way to learn more about the world of bugs, but what does that mean for our privacy when we may not even be aware of the camera? As our devices get smarter, it’s getting harder to recognize when our data is being used, where, and for what purpose. Rokid, a technology company specializing in robotics and AI, recently developed temperature-reading glasses that allow wearers to know if they’re coming close to anyone with a fever.21 This may be great for social distancing and minimizing the risk of catching coronavirus, but now the personal body temperature of anyone who crosses the line of sight of the wearer can be accessed with no opportunity for consent. Drone surveillance has come under fire after a Predator drone was seen flying over protests in Minneapolis in May; even smaller drones only heighten the concern that our movements and behaviors can more easily be monitored.22 Will the public really trust tech companies with our brains’ data when we’ve seen what they’ve done with our much less personal information?
Security. With so many gadgets entering our homes, workplaces, and even our bodies, we have more potential for vulnerabilities. Many of our gadgets can now listen to our conversations or see us without us even being aware. A $1,300 smart crib was discovered to be vulnerable to a hack that would rapidly rock babies back and forth.23 Last year, the FDA warned that some Internet-connected insulin pumps had vulnerabilities that could be exploited to over-deliver insulin or stop insulin delivery altogether.24
Ethics. Adding AI to our multitude of devices may sound like a panacea – who wouldn’t want a lawnmower that could recognize the difference between a rose and a weed?25 – but we can’t forget that we still haven't figured out how to use AI responsibly. Example after example of biased AI systems that can deny you a loan, a job, and even have you arrested serve as sobering reminders that we are not ready for this level of AI.26 Consider also that many of these tiny technologies are not cheap. For example, the Rokid temperature-reading glasses mentioned above currently cost over $7,000, potentially exacerbating already unacceptable disparities between the digital haves and have nots.
2 Li, J., Thiele, S., Quirk, B. C., Kirk, R. W., Verjans, J. W., Akers, E., ... & McLaughlin, R. A. (2020). Ultrathin monolithic 3D printed optical coherence tomography endoscopy for preclinical and clinical use. Light: Science & Applications, 9(1), 1-10. https://doi.org/10.1038/s41377-020-00365-w
19 Udugama, B., Kadhiresan, P., & Chan, W. C. (2020). Tunable and precise miniature lithium heater for point-of-care applications. Proceedings of the National Academy of Sciences, 117(9), 4632-4641. https://doi.org/10.1073/pnas.1916562117