#NEXTalks: The AI Oligarchy
Written by: Dr. Foteini Agrafioti
Dr. Agrafioti is the Chief Science Officer at RBC and Head of RBC Research Institute. She is responsible for RBC’s intellectual property portfolio in the fields of artificial intelligence and machine learning. Prior to that, she founded and served as CTO of Nymi, a biometrics security company and maker of the Nymi wristband. Foteini is an inventor of HeartID, the first biometric technology to authenticate users based on their unique cardiac rhythms. She is also a TED speaker and serves on the editorial review boards of several scientific journals. Foteini was named “Inventor of the Year” in 2012 by the University of Toronto where she received her PhD in Electrical and Computer Engineering. She is an alumna of the Next Founders, one of three entrepreneurial programs delivered by NEXT Canada.
We tend to throw the word “disruption” around a bit too loosely in tech. Like the boy who cried wolf, there’s risk in not paying attention when something truly disruptive comes around.
If artificial intelligence delivers on its promises, it will be a force that can move society forward. This is not an understatement. With applications across healthcare, climate, transportation, financial services, education, and more, AI has the potential to become a fundamental element of human society, and ours a rare generation in history that gets to live such a seismic transition.
Big technology companies have already built fortified empires around machine learning. Critical intellectual property, talent and capability has concentrated in a handful of tech giants. While this ensures that the field is well funded, a question naturally rises:
Does this domination of AI by a few corporations move the world forward or not?
I believe not. First, I want to be clear that I have nothing against the notion of free market and I respect every company’s responsibility to capitalize on opportunities. However, I can’t avoid considering the implications of what could happen if AI continues to be brought to market by a small handful of entities.
The path of an enterprise is naturally focused on its strengths and, when it comes to machine learning, strengths are defined by readily available data. In the simplest case, one can consider the opportunities to benefit our society that may not materialize because they weren’t on the critical path for these few players. And the items that were on that critical path will completely upend their respective markets, leaving them exclusively owned and monopolized.
It used to be the case that scientists in universities and research centres solved fundamental problems for the world and transferred that knowledge to the industry. Today, select problems are solved within the few companies that have the capability to do so. In fact, the best place for a young scientist to learn machine learning has altogether ceased to be the university. A handful of companies worldwide have become the go-to places for aspiring students to get educated in the state-of-the-art, and this is a direct result of the numerous academic luminaries that have chosen to move their labs there.
Industrial R&D centres have shifted their focus from trade secrets to open academic practices such as scientific publications of algorithms and results. Part of that shift, no doubt, is in response to pressure from the academics that have overtaken these industrial R&D centres. While open access to the research is a good thing, it’s important to note that these publications rarely contain the necessary ingredients to commercialize the technologies that are released to the public, nor are they easily replicated by those who lack the means (talent, processing power, datasets, etc.). They’re essentially a smokescreen, artfully advertising their wares while fogging up the marketplace window. In simpler terms this means our collective understanding of the science may be advancing, but our ability to commercialize these technologies and solve problems for our societies remains limited.
That is why it’s crucial to diversify our focus. In addition to the work that is done within large technology companies, we must create opportunities for R&D work in non-traditional AI playgrounds (enterprises or startups) and those at the forefront of the AI field should consider the role they play in breaking the AI oligarchy. Because if they don’t, there is innovation — potentially vital and life-enhancing innovation — that may never see the light of day.
Two great examples of such innovations (that wouldn’t make financial sense for a tech giant to develop) are currently coming out of NextAI and Next 36, two in a series of Canadian entrepreneurship programs run by NEXT Canada.
Intuitive Inc., a BC-based start-up is using machine learning to create “smart bins” that automatically sort our garbage into recyclables and non-recyclables. The bins use image recognition to identify objects and separate them into the right container based on a city’s municipal guidelines. With urban populations alone generating 1.3 billion tons of waste per year and only 2 per cent of recyclable materials currently making it to recycling plants, there’s wide-ranging environmental impact.o
And Orbityl, a lab founded by two McGill engineering graduates, has created a pair of earplugs embedded with machine learning algorithms to track the electrical activity in the brain of insomniacs. The idea is to provide them with a customized sleep plan based on their unique habits and patterns. The data collected by this technology can subsequently help researchers find solutions to treat the myriad health problems related to sleep disorders.
Both these companies — and hundreds like them — are using AI to build products that could trigger a windfall of positive benefits to humanity. But for those benefits to be realized, these companies need to exist, grow and be backed by AI experts and funders.
We can change this course by supporting a diverse ecosystem that allows our AI community to stick to what they’re doing and feel that they can continue to do their best work outside the orbit of tech giants. Unlike prior socioeconomic revolutions, it is entirely within our power to determine how the future of our relationship to AI will play out, but the speed at which the technology is moving demands immediate action. Let’s program it the right way.
NEXTalks features thoughts on entrepreneurship, innovation and giving from the NEXT Canada network. The goal of the NEXTalks series is to spark dialogue around important issues related to innovation and technology.