Of the start-up and AI researchers disconnected from the real world? The tens of millions of dollars invested by Quebec and Ottawa in artificial intelligence over the last five years have not produced the hoped-for results, on the contrary. In health, the results are a failure, mixed with conflicts of interest in the management of public funds.
This is the conclusion reached by researcher Myriam Lavoie-Moore, in a report published Thursday morning by the Institute for Socio-Economic Research and Information (IRIS). “In 2018, we were already questioning the Canadian AI strategy. By targeting health more precisely, I found a series of problems that emerge as much in corporate ethics as in a misalignment between investments and real public health problems,” summarizes the IRIS researcher in an interview. with The duty.
This assertion is based on the inventory of some 312 AI projects which have received government funding over the last five years, and on which data has been made public.
IRIS’s first observation: despite this large number of subsidies, only “a handful of AI companies have benefited from them”, and even there, we don’t really know where all the money went. IRIS cites as an example MEDTEQ+, a consortium which, according to the research institute, has received more than 63 million from the Quebec Ministry of Economy, Innovation and Energy in five years. IRIS was only able to find a fraction of this amount, i.e. 8.8 million.
IRIS also observes that MEDTEQ+ has included on its board of directors in recent years directors of companies which are among those which have received the most public funding, including Imagia and AlayaCare. “The ministry distributes subsidies to unaccountable companies. While we constantly talk about ethics in AI, there are significant blind spots when it comes to its financing,” says Myriam Lavoie-Moore.
The Imagia case
Another observation from IRIS: AI funded by the governments of Quebec and Canada is not used to solve the most urgent public health problems, such as access to professionals, administrative burden or the famous digital medical record . “We are really not meeting the needs of the sector,” says the researcher, who adds: “In fact, we are helping companies that accelerate the privatization of health — for example, telemedicine services offered in insurance programs private companies that divert public resources. It’s quite scandalous. »
IRIS cites a few companies that illustrate this disconnect between AI research, technology funding, and their lack of success both in helping public health and in finding any commercial success. The most obvious is the Montreal medical imaging specialist Imagia. The Mile-Ex company founded in 2015 thought it had a revolutionary AI application in its hands: its algorithms could accelerate the analysis of x-rays to detect cancer cells with an unprecedented degree of success.
Despite the promise of its technology, Imagia never managed to establish itself commercially. In February 2022, the company announced its merger with the medical company Canexia Health, from Vancouver, and completed a new round of financing of 20 million, notably from the BDC. Then, in June 2022, Imagia Canexia Health obtained 18 million more from a federal program.
A year later, last August, Imagia Canexia Health declared bankruptcy. No one involved with the company wants to comment publicly on this debacle.
“The example of Imagia, in my opinion, is the best illustration of the disconnection from AI funding,” says Myriam Lavoie-Moore. The technology was very promising, with the potential to create a Quebec unicorn. But for the health network, it makes no sense. The health system does not block at the level of the analysis of the images, it is well before that that it blocks, at the level of the consultation and the prescription, of the image taking, then at the time of applying the treatment. This demonstrates the misalignment between the direction of public funding and the needs of the system. »
Failure in AI
The IRIS report on the financing of AI applied to the health sector is not the first of its kind to speak of strategic failure. Last year, it was the Institute for Research in Contemporary Economy (IREC) which said essentially the same thing, with one nuance: IREC found that public money generally ended up in the pockets of foreign companies, to most of them already well off.
The $1.17 billion invested in AI by governments between 2016 and 2020 created companies and technologies that were quickly acquired by rival companies based outside the country, which should have raised serious questions about the effectiveness” of the method of financing AI in Canada, concluded the IREC, which itself echoed a similar publication from the Canadian Council of Innovators. This think tank founded by Ontario billionaire Jim Balsillie declared in 2018: “For too long, the repercussions of these policies have been negative in terms of the ability to transform local companies into global giants. »
All this further reinforces the conclusion of Myriam Lavoie-Moore, of IRIS: “We must not reject technology altogether, but we must certainly better manage and supervise the financing of innovation in the country, then develop strategies with the people who will be impacted. »