Artificial intelligence must be put at the service of society’s challenges. This is, in essence, the course set by tech leaders at COP26 last year. At a time when societal challenges are piling up, this ambition is obviously laudable. However, it raises the issue of runoff. Indeed, if the Gafa, the “first on the rope” in tech, have the means and the (displayed) desire to domesticate AI, what about other companies? How to convey this awareness to the latter and give them the means to use it? A global challenge for the next ten months.
Contrary to what one might think, the conditions for this runoff are far from being met in Europe. Admittedly, a regulation has been emerging since last spring, but it has two major limitations.
Disconnection with reality
The first is philosophical. This regulation, in fact, is more of a deterrent than an incentive. For example, if a risky artificial intelligence system is placed on the market, the fine may reach 30 million euros or 6% of the company’s annual worldwide turnover. Compliance, on the other hand, is not rewarded, contrary to what happens for example in the automobile industry, with the CAFE (Corporate Average Fuel Economy) standard.
The second limit of the European project is its disconnection from reality. It sets an objective, while being careful not to give companies the keys to achieving it. In addition, let’s not lose sight of the fact that compliance with regulatory frameworks requires resources that strain project budgets.
Unleash Artificial Intelligence
In terms of AI, the reality of companies is laborious. There is, on the one hand, the 25% at ease with artificial intelligence and on the other, the 75% for whom the appropriation and industrialization of this innovation are still problematic. These organizations always come up against obstacles such as business value, change management, and even distrust of this technology expressed by end customers. As long as these blockages last, it will remain perceived as a privilege.
The notion of ‘common’ must be put at the heart of artificial intelligence if we wish to emancipate it from the obstacles that constrain it within companies.
In 1983, Richard Stallman, then a young computer scientist at MIT, laid the foundations of the free software movement to which we owe Wikipedia, Mozilla and Creative Commons licenses. Forty years later, this notion of “common” must be put at the heart of artificial intelligence if we wish to emancipate it from the obstacles that constrain it within companies.
Take the step of industrialization
“Debiasing” techniques, analyzes of data sets (data sets) or modeling choices must become “commons” accessible to data scientists from all organizations in order to take the industrialization stage.
Once this phase has passed, companies will then be able to think “sustainably” and equip themselves with tools and processes, tools for monitoring solutions based on a predictive model, impact analysis methods for the use of personal data, to precisely systematize the sustainability of their projects.
Through their CO emissions2, companies obviously have a role to play in the face of climate change. Problem: only 9% are able to measure their CO emissions2. A very low figure that the science of the data would however make it possible to raise. Through its ability to measure, analyze, and trigger action, data science is a key lever in the ecological transition of companies. Contemporary climatic and ethical issues require its emergence and generalization. But if it is not made available to companies, it will remain only a concept.
Laurent Felix is Managing Director France of Ekimetrics
We would like to thank the writer of this short article for this incredible material
opinion | AI: connecting discourse and business reality
Explore our social media profiles along with other related pageshttps://www.ai-magazine.com/related-pages/