A framework for Artificial Intelligence in the enterprise

The article by Frédérique Roseau, editor-in-chief of ©RevueFiduciaire – RF Social n° 258 – January, highlights the crucial issues raised by Jennifer Carrel and Anne-Marie Pecoraro, partners at UGGC Avocats.

AI with risks

“The fact that employees use AI without the employer’s knowledge presents many risks: breaches of the RGPD and of personal and/or strategic data… There may be leaks of information damaging both to the company and to its customers to whom it guarantees in particular the confidentiality of exchanged data and the protection of intellectual property (this is all the more important in the creative industries). However, it is also possible that the use of AI may not bring added value but, on the contrary, a loss of value (e.g.: reproduction of what already exists without any original contribution).

From a human resources point of view, the use of AI must not lead the company to contravene the rules laid down in the French Labor Code. For example, its use in a recruitment process must not be at the expense of the principle of non-discrimination. Indeed, artificial intelligence can have biases that human intelligence must correct.

So it’s important to focus on how employees use AI.”

An indispensable framework

“There are several ways of legally framing the use of AI by employees. The first is the employment contract, and we certainly need to review the clauses in use to adjust them to AI (e.g.: reinforced discretion clause). However, using the contract, which is rather fixed (its modification presupposes the employee’s agreement) whereas AI is evolutionary, is not the most suitable solution.

Setting up an AI user charter is the most logical and progressive approach. The employer must ensure that it is adopted in accordance with the internal regulations process, notably by consulting the CSE, informing the labor inspector, who will retract any disproportionate infringement of employees’ rights, and informing employees of its existence so that it can be enforced against them. Consequently, if all the conditions are met, an employee who uses AI without respecting the framework laid down by the charter could be sanctioned.

In addition, it’s important for employers to train employees in the use of AI, as part of their obligation to train and adapt to modern means of working.It’s therefore difficult for companies to pretend that nothing is happening; they have to take action to protect themselves. What’s more, European regulations will eventually strengthen organizations’ obligations in the field of AI (European regulation 2024/1689 of June 13, 2024).”