2023-12-19 11:20:00
Only 4% of French companies are fully prepared to integrate artificial intelligence, revealed in November a recent study published by the American technology group Cisco. And yet, all companies, whatever their size or sector of activity, will integrate AI into their organization tomorrow.
Already full members of our daily lives, from Alexa to Siri, so-called weak AIs, according to the definition given by Elisa Fromont, professor at the University of Rennes, can thus take the form of a predictive model used to estimate risks for insurance or detect illnesses. With weak AI, “ the machine must arrive at the same solutions as humans, regardless of the method, for targeted tasks, requiring reasoning, learning, or decision-making » said the academic during the Imagine Summit organized on December 7 in Rennes.
However, for businesses, the arrival of generative technologies is accelerating things and their applications offer much more advanced possibilities than ChatGPT. As part of a fictitious AI trial, organized on December 6 at the IRT b-com in Rennes, the Ille-et-Vilaine CCI wanted to question the place that this technology will take tomorrow and its possible capacity to run a business. A big leap into the future to better understand the impact, challenges and risks of AI on organizations.
A defendant, lawyers, experts
Close your eyes, it’s January 2036: the line between human and machine has blurred, and Bruno’s day is driven by artificial intelligence and autonomous robots. Like the other employees of Mobility Less Carbon, a leading company in carbon-free urban mobility, this R&D engineer learns that the manager, who is seriously ill, has handed over her company to Emma Thompson, an artificial intelligence, currently director of production. Worried, the employees decided to file a “ class action complaint for discrimination once morest human beings ».
The trial of Paule Le Gall opens on June 4, 2036 at the European AI Tribunal on the basis of three charges: discrimination, disloyalty, transhumanism for having given the power to an American AI to “ perpetuate your conscience “. The exchanges between the lawyers, the experts and the witnesses of the defense and the civil party called to the bar, begin, under the eye of jurors called to condemn Paule Le Gall or not.
« There were buyers, Americans, Chinese, Russians, would you transfer a company to Russians today? » asks the latter. The accused explains that she decided to create Emma Thompson, a human projection in her image, then to sell her business to her, following meeting Mia, an AI at the head of another prosperous Breton company.
Limitless automation, human and climatic costs
Furthering reflection on the role of AI in industrial processes (automation, valorization of high value-added tasks, etc.), in the invention of new professions of the future (for example that of ambassador of human cohabitation and machines in the professional environment), in the sense given to work (technicians transformed into robot assistants), this true-false trial also explored the social and psychological consequences (loss of confidence, anxiety, etc.) of arrival of AI within Mobility Less Carbon.
Described without empathy, without human emotion in its working relationships and the management of a company, a managerial AI cannot, according to an employee, be a reflection of its employees or their values.
« By believing her employees to be incapable of taking over her business, Madame Le Gall considers them idiots » judge for his part Cédric Sauviat of the French association once morest AI (AFCIA) adding that “ the quest for automation has no limits.” Another specialist returns to the climate cost of version 17 of ChatGP: the equivalent of around ten aircraft carriers in terms of carbon emissions. On a human level, an AI “ having no conscience » might not adapt to all contexts of changes in the payroll. Depending on the data that trains Emma Thompson, she may not recognize a non-white employee or speak a language that has not been integrated into her.
Ultimately, it is the human who holds the tool
« I am not the thing, my creator wanted me to integrate into the world and be able to do what humans must do but much more like finding solutions to fight global warming » however defends Emma Thompson, supported by Stéphane Paquelet. According to the head of the b-com AI laboratory, the action of this very self-confident robot with a very human image remains “ very supervised » by French and European legislation and by its board of directors.
« This technology-assisted sharing of powers harms neither the interests of the company nor the general interest represented by the law » he assures. In terms of security and authentication, AI productions can also be guaranteed by a signature mechanism. It is the human who holds the tool, confirms another expert, sweeping away any risk of interference from a competing country.
At the end of the discussions, the prosecutor’s requisition and the lawyers’ pleadings, the public rendered a judgment which we will not reveal. Organized in particular with the Rennes Bar Association, this immersive true-false trial is now accessible online on YouTube. Virtual but informative, it is part of an acculturation program to artificial intelligence that the Ille-et-Vilaine CCI will launch in early 2024.
1702987378
#Rennes #scenario #company #run #generative #scrutinized
