Canada-UK Colloquium on AI & Society
Last fall I attended the Canada-UK Colloquium on AI & Society, held at the Munk School of Global Affairs and Public Policy. The three-day colloquium gathered experts from academia, industry, government, etc. from both countries, to discuss the opportunities and challenges brought upon by the increased use of machine learning and algorithms in our societies. The participants were asked to propose specific policy recommendations for our respective governments.
As the official rapporteur, I was tasked with producing a report that encapsulated the discussions and recommendations. (This also answers the question: “what did you do during your Christmas holiday?”) The meeting was held under the Chatham House Rule, which means no statement is attributed to a speaker or their affiliation.
The full report can be accessed via the link below. Some interesting points, lifted directly from the conclusions section, are:
We are at the foothills of realizing AI technologies’ full potential, with some rapid advances in recent years and early adopters in pioneer sectors indicating the breadth of which AI could transform our society and economies. Currently, most of what we consider today as AI is actually a weak form involving machine learning and deep learning applied to specific tasks. As we plan for an AI-powered future, we must also prepare for strong or general AI where a single AI solution is performing many tasks and learning as it works.
It is important to continue to support the other sciences and technologies that are essential for AI to succeed (sensors, batteries, dynamics user interfaces, cloud services, data management etc.).
A good AI “ecosystem hub consists of start-ups and small and medium companies working alongside universities hand large tech companies that to act as “anchors”. These tech companies attract talent and can take on the risk of translating academic discoveries into useable software. Governments need to support the establishment of these AI hubs and actively remove inhibitors that prevent small businesses in these hubs from scaling-up globally.
Public engagement is very important. The public and private sectors and academia need to be open and honest with the public about the opportunities and trade-offs of adopting AI and automation. This will help allay fears of automation and AI and ameliorate the rise of populism and nativism.
We need to clarify and use the proper terminology when talking about AI. Government, industry, academics, journalists, etc. should use language that is accurate, clear, specific, consistent, and simple for the public to understand.
Governments should be cautious not to over-regulate, which ends up becoming a “tax” on corporations to do business. Small companies may not have sufficient resources to address complex and possibly conflicting laws from many jurisdictions. As well, some technologies are covered by existing laws, rules, regulations, and mechanisms. Of course, regulation can also be positive for innovation but there needs to be a balance.
CUKC 2018 Report
(Photo credit: Milan Ilnyckyj)
