Artificial intelligence arouses fears and hopes. The automation of certain tasks could eliminate 300 million white-collar jobs in the coming decades. Law seems to be one of the professional fields that could be the most affected. At the same time, new analytical capacities are already offering new ways to tackle certain social problems, such as school dropout. Welcome to your near future…
As soon as the 4e version of the chatbot ChatGPT went online in mid-March, cardiologist Anil K. Gehi, a professor at the University of North Carolina at Chapel Hill, decided to test its revised and improved analytical capabilities. For this, he asked for a diagnosis from the medical file of a patient admitted the day before to his hospital department. The list of complex symptoms included several specialized terms, but the question posed was very simple: how should this patient be treated?
ChatGPT-4, a robot from the OpenAI company, offered exactly the same treatment as that adopted by medical teams. Professor Gehi repeated the experiment. Each time, the robot fell exactly on the same answer as the caregivers.
“The robot still needs experts like the Dr Gehi to judge his answers and perform the medical procedures, commented on the New York Times recounting this experience. But he can exhibit that kind of expertise in many areas, from computer programming to accounting. »
Robotization developed at the end of the 20th centurye century has affected jobs blue-collar workers, manual jobs. In an automotive assembly line, robot carriers or welders very effectively replace workers for repetitive and alienating tasks. The new ones chatbotsable to discourse on any subject in natural language, will now especially destabilize white-collar jobs.
The fighting machine
In short, the digital revolution continues and its effects in the world of work should increase with this new stage of conversational robots, that of Web 4.0.
“Tools appear all the time. With conversational robots, we are seeing a deeper advance that should not be overlooked because they are seeking skills that were still reserved for humans until now,” explains Xavier Parent-Rocheleau, professor in the Department of Human Resources Management. at HEC, specialist in artificial intelligence and digital technologies.
“The liberal professions are based on knowledge acquired after long years of study. The public pays for access to this proprietary knowledge. AI propels what started with search engines by democratizing access to knowledge. The exclusivity of the possession of knowledge is weakening. The price of knowledge and services falls. »
He cites as examples online brokerage operations, platforms for filling out tax returns, language or training tutorials. Likewise, certain aspects of the legal or medical professions can and certainly will be automated. “It’s completely correct because in some aspects, the machine is much more efficient,” says Professor Parent-Rocheleau.
He also recalls that the book The Future of the Professionsby Oxford professors (father and son) Richard and Daniel Susskind, already announced in 2015 (a new expanded edition has just been published) the massive technological layoff of doctors, accountants, teachers and members of the clergy.
A study from the University of Pennsylvania unveiled a few days ago, funded by OpenResearch, predicts that these new digital tools will especially disrupt better-paid jobs that generally require university levels of education, up to the doctorate. The survey suggests that half of the tasks of 20% of jobs in the United States will be automated by the new wave. Above all, the study announces a great replacement for professions dealing with information, the great strength of AI. The list of the most transformed, even threatened, includes legal clerks, pollsters, accountants, financial analysts, mathematicians, translators, proofreaders, writers, public relations officers and…journalists.
A report by investment bank Goldman Sachs released this week predicts that artificial intelligence could replace 300 million jobs in the West, or a quarter of the long-term workforce. Projected disappearances include 46% of administrative jobs and 44% in law-related occupations, pointing out that salaries for the remaining jobs in these sectors are also expected to fall.
Technology, on the other hand, could create new jobs. The usual routine, what. Six out of ten jobs that currently exist did not exist in 1940. The great change to come should not, however, affect service jobs such as mechanics, cooks, bartenders, plumbers or dishwashers. It’s the revenge of the textbooks…
Orders and disorder
A working group bringing together academics and delegates from professional orders in the health sector in Quebec studied the impact of AI, including by surveying the members of 25 recognized corporations. The analysis published in 2019 shows that artificial intelligence systems are already used in common practices which are in great need of regulatory guidelines.
“For several years, we have seen that certain AI systems are more efficient than panels of professionals of excellent reputation, for example in detecting various ocular pathologies in certain patients. It’s very challenging “says Marco Laverdière, Director General of the Order of Optometrists of Quebec, who co-conducted this survey of health professional orders representing hundreds of thousands of professionals.
“We are at the stage of asking ourselves how we are going to live with this transformation in each of the orders. Concerns are experienced in different ways and, yes, it is possible that certain professions will disappear or in any case be radically transformed, even if the forecasts are more about AI help in decision-making. »
He illustrates his point by evoking the problem of comorbidities, which are so important in the risks of developing lethal diseases. We have seen this during the pandemic. AI can well determine probabilities and suggest interventions. In fact, each person treated will benefit from a human encounter that can refine the understanding of their particular physical and mental state. “It seems to me that we are more at the stage of deciding how to integrate AI into professional practices and according to what criteria,” summarizes Mr. Laverdière.
The risks seem as huge as the benefits. Poorly designed systems could lead to fatal errors or introduce interpretive biases that add discrimination for certain people or social categories. The capture of public and personal data by large techno-industrial groups is already arousing innumerable socio-political fears.
courage against the machine
The third part of the federal government’s Bill C-27, before the House of Commons, enacts the Artificial Intelligence and Data Act (LIAD) to “regulate international trade and commerce” and to “mitigate the risks of harms and biased outcomes associated with high impact AI systems.” The Quebec Institute of Artificial Intelligence (MILA) and UNESCO have just launched in Montreal the collective work Artificial Intelligence Governance Blind Spots on the need to regulate these unprecedented technological changes.
The 2019 survey of health professionals also identified specific issues in terms of training, protection of personal information, quality and safety of care and the power of intervention of the orders.
A prototype code of ethics was proposed by Mr. Laverdière and his colleague Catherine Régis, from UdeM, to support the integration of AI into professional health practices. The ten-point proposal is inspired by the Montreal Declaration for Responsible AI Development.
The first principle affirms that the professional must ensure that he is adequately trained to use the new tools, in particular to know their possibilities and limits. The second affirms that the professional must inform the patient about the use of the information collected by the AI and respect his choices.
“The goal of the exercise is to move from the main principles and major concerns to a form of concrete framework for the integration of AI into professional practices, explains the professor who hopes to shift to an adoptable version. code by the middle of the decade. But it is not completely abnormal that the law is behind on this kind of phenomena. It is even desirable in some cases so as not to kill innovation based on false fears or beliefs,” notes Mr. Laverdière.