Industry-wide ethical standards for artificial intelligence (AI) may currently be lacking but the CIO Strategy Council says it is ready to create standardized policies to help Canadian companies and governments.
After receiving input from across the country, the council’s standards committee announced in a meeting last week that it has refined and approved the policies it sees as most important to creating ethical AI standards policies, Jenna Sudds executive director of the CIO Strategy Council told IT World Canada.
With this announcement, two technical committees are currently being set up and will include industry leaders from across private and public sectors, who will be tasked with creating the actual standards documents.
The committees will focus on two topics; data access and privacy as well as ethical design and use of automated decision systems.
The CIO Strategy Council, as IT World Canada has previously reported, which was co-founded by the Chief Information Officer for the Government of Canada Alex Benay, aims to be a leader in developing the first-ever globally recognized standards for the ethical use of AI and big data.
“This work really stems from this perceived gap in Canadian participation as well as international standards leadership with respects to ethical AI,” says Sudds, “We saw this as an opportunity to lead and get ahead of that race as a country, not just government, but also as citizens.” She sees the council’s development of standardized policies as a way for Canada to be a leader in ethical AI both nationally and on the international stage.
The policy documents that the council is creating cover a wide range of topics but focus mainly on how organizations can develop ethical automated systems and how those AI systems can meet privacy requirements when collecting and using information.
The committee on the design and use of AI systems will address topics like; unintended biases in algorithms, ‘black boxes’ (the unknown inner workings of a system’s algorithms) and whether that information should be made public, and the adoption of Algorithm Impact Assessment frameworks.
The committee on data access and privacy will look at a range of topics that include; control and ownership of data, consent requirements, and data aggregation and storage.
Sudds says that if done well these standards will give the Canadian technology industry a competitive advantage.
She says that the input and open discussions that the council is fostering from both private and public participants, will create standards that will represent Canada’s strong contributions, on an international stage, in the world of AI.
The Canadian government itself is currently working on its Pan-Canadian Artificial Intelligence Strategy, dividing $125 million in funding between three prominent Canadian AI institutes, with one of its goals to develop ideas around the economic, ethical, policy and legal implications of advancements in AI.
And while Sudds agrees this move by the federal government is a step in the right direction, she says that the consensus at the council is that, “we haven’t done nearly enough, we are just at the tip of the iceberg.”
“Right now it’s the ‘wild west’ and its scary times but the good news is that work is happening and it’s just a matter of putting that work together and getting Canadian’s engaged,” says Sudds.