Over a year ago, the federal government published a voluntary Code of Conduct for the responsible development and management of advanced generative AI systems1
.
While these systems offer many benefits, such as writing e-mails, answering complex questions or producing realistic images or videos, they carry significant risks to
health and safety, propagate prejudice and have wider societal repercussions, particularly when used by malicious perpetrators.
It's telling that even the pioneers of artificial intelligence are wary of the risks it presents. In early 2023, over 350 AI industry leaders signed a declaration2 urging the
international community to make it a priority to mitigate the “extinction” risk posed by AI, placing it on a par with a pandemic or nuclear war. Since this moratorium has
not yet taken place, the urgency to act is arguably even greater, as Joshua Bengio, one of the signatories of the moratorium request, expressed in an interview with
Les Affaires3
.
Conscious of the importance of a proper framework, and in order to manage and mitigate these risks, the federal government invited companies to become signatories
to this Code, which describes:
- the measures that should be applied pending the adoption of regulations under the Artificial Intelligence Act;
- as well as additional measures that should be taken by any company that develops or manages the operations of such a system made accessible to
a wide public, i.e. systems with a wider range of potentially harmful or inappropriate uses.
The purpose of this proposal is to invite the Board of Directors to adhere to this Code of Conduct.