Yoshua Bengio, a recognized trailblazer in the field of artificial intelligence, has openly expressed his concern about the unanticipated rapid progression of AI technology and the failure of Canadian legislation to keep pace. Bengio, a pivotal figure in AI innovation and head of the Mila AI research institute in Quebec, recently shared his disappointment about the lumbering progress of Bill C-27.
Bill C-27, officially known as the Digital Charter Implementation Act 2022, comprises the Artificial Intelligence and Data Act. This act is designed to set up regulatory boundaries around the use of AI technology. Despite initial progress earlier this year, the bill is currently stalled in review.
“I was hopeful, six months ago, that this prospective law would have been in effect as we speak. However, I am unsure if that will be realized as the legislative process can often be painstakingly slow,” Bengio voiced.
Expressing his belief in the urgency of swift governmental action, Bengio acknowledges the breakneck speed at which the technology has grown, surpassing even his own projections. This was further exemplified by the launch of ChatGPT, a humanlike AI chatbot by San Francisco-based OpenAI last November.
This revolutionary communication AI spurred a tech competition involving technological titans Google and a fleet of venturesome startups. The technology has already seen new developments, including voice and image integrative capabilities.
“We find ourselves coursing a path laden with uncertainty, while it is evident that the power of AI systems is steadily increasing. The technology may eventually mirror human abilities in numerous areas,” Bengio noted.
Understandably, the potential for such advanced AI is exhilarating, but also alarming due to its potential for misuse. This concern prompted Bengio and fellow AI pioneer Geoffrey Hinton, to call attention to the dangers of the unbridled use of this technology.
On Tuesday, the duo amplified their message, issuing a document penned by approximately 20 other AI experts, appealing to governments to stringently control the risks associated with AI.
To outpace the evolving technology and minimize risks, they insist that governments and businesses allocate a third of their AI research and development budgets toward the safe and ethical implementation of AI technology.
Additionally, they are lobbying for AI creators to invest in safety measures that can curb their products’ capabilities and for companies to bear responsibility for any harm their technologies may cause.
Acting swiftly is pivotal, they contend, as AI has already shown capabilities of surpassing some human abilities, posing threats that could magnify global inequalities, propagate automated warfare, manipulate public sentiment, and intensify surveillance.
So far, humanity finds itself immersed in a disorderly, unregulated competition, Bengio lamented. “Government and the general public’s grasp of AI is minimal, which needs rapid change,” he expressed.
Hope for change could be on the horizon with Bill C-27, which if enacted, would mandate businesses to elucidate their reasons for AI development and appoint a independent commissioner for compliance audits regarding AI regulation.
In the interim, a voluntary code is being employed, endorsed by several corporate stalwarts in AI, with the exception of Shopify Inc. whose founder, Tobi Lutke, declined participation.
On addressing reservations about regulatory barriers, Bengio asserts that it is crucial to expedite the process to level the field for all companies. “It would be inequitable if the conscientious companies bear the responsibility to protect the public while others exploit the situation to secure a larger market share,” he said.
“Self-regulation assists in navigating these dire straits, and may act as the interim solution. However, it is essential that safety and ethical norms find a permanent seat in all company policies,” Bengio concluded.
