ChatGPT burst onto the public scene late last year, giving artificial intelligence its “aha moment” for many people. AI is now seemingly everywhere, attracting enormous attention and excitement alongside concerns, legal threats and talk of regulation. The potential of AI is evident to just about everyone, but the challenges associated with bias, copyright, privacy, misinformation and more can’t be ignored. Cohere AI is a Canadian-based AI firm that is widely viewed as one of Canada’s AI stars for its large language models that enable companies of all sizes to integrate AI technologies. Aidan Gomez, who worked on the “T” in ChatGPT, is the co-founder and CEO of Cohere AI. He joins the Law Bytes podcast to talk about AI and his views on the myriad of emerging legal and regulatory issues.
Articles by: Michael Geist
Bill C-11 Estimates Revealed: Internal Government Documents Show No Impact on Net Employment, Admit Streamers Already Invest Millions in “Unofficial Cancon”
The government’s support for Bill C-11 has often been framed on economic terms with Canadian Heritage Minister Pablo Rodriguez arguing that the bill will “create good jobs for Canadians in the cultural sector”. I’ve long maintained the government’s claims that the bill will generate billions of dollars in new money was massively exaggerated and that a far more likely scenario would be that the bill would simply lead to a reshuffling of existing expenditures.
Using the Access to Information Act, I have now obtained a copy of the government’s internal estimates for the economic and production impact of Bill C-11 (methodology, memorandum, PPT), which confirm many of my suspicions. While the government is pinning its hopes on massive spending from Internet streamers such as Netflix, it admits that even if the bill did not pass it would not affect net new employment in the sector. Moreover, internally the government recognizes the claim that Netflix and foreign streamers don’t contribute to Canadian content is false, as it has identifies a new category of “unofficial Cancon” which would qualify as Cancon under every measure but for the fact that it is owned by companies like Netflix and Disney. And as for the payments from social media companies that the government insists are so essential that it has fought for years to include user content regulation in the bill? The estimated economic benefit represents just one percent of its total projection for Bill C-11 with pure guesswork about what percentage of content on the platforms might require contributions.
Government Departments Pressure Social Media Sites to Censor News Links, Mean Tweets
The risks associated with the government’s online harms (or online safety) plans is not limited to Canadian Heritage’s credibility gap, which as I’ve recounted has included omitting key information in its public reports on consultations and shocking efforts to exclude contrary voices altogether. A new report, based on the government’s response to a Parliamentary Written Question from Conservative MP Dean Allison (Sessional Paper No. 8555-441-1219) raises new concerns about efforts to censor social media. The written question asked the government for “requests made by the government to social media companies to take down, edit, ban or change in any other way social media content, posts or accounts since January 1, 2020.”
The resulting report is stunning as government departments have in fact pressured the social media companies to remove news links and a range of lawful content including tweets that are said to contain “offensive language” or an “offensive reply.” The use of government power to censor social media posts or news links is exceptionally dangerous and crystallizes the fears of regulatory powers without the necessary due process and oversight that could lead to censorship of lawful content online.
The Canadian Heritage Online Harms Credibility Gap, Part Two: Filtering Out Critics From Participating in Anti-Hate Consultation Survey
The government’s online harms bill, led by Canadian Heritage, is likely to be introduced in the coming weeks. My series on why the department faces a significant credibility gap on the issue opened with a look at its misleading and secretive approach to the 2021 online harms consultation, including its decision to only disclose public submissions when compelled to do so by law and releasing a misleading “What We Heard” report that omitted crucial information. Today’s post focuses on another Canadian Heritage consultation which occurred months later on proposed anti-hate plans. As the National Post reported earlier this year, after the consultation launched, officials became alarmed when responses criticizing the plan and questioning government priorities began to emerge. The solution? The department remarkably decided to filter out the critics from participating in the consultation by adding a new question that short-circuited it for anyone who responded that they did not think anti-hate measures should be a top government priority.
The Canadian Heritage Credibility Gap on Online Harms, Part One: Public Report Did Not Disclose 90% Opposition to Its 2021 Proposal
The government’s online harms bill – likely rebranded as an online safety bill – is expected to be tabled by Canadian Heritage Minister Pablo Rodriguez in the coming weeks. The bill, which reports suggest will even include age verification requirements that raise significant privacy and expression concerns, is expected to emerge as the most controversial of the government’s three-part Internet regulation plan that also includes Bill C-11 and Bill C-18. Given the fierce debate and opposition to those two bills, it may be hard to believe that online harms or safety will be even more contentious. Yet that is likely both because the bill will have enormous implications for freedom of expression and because Canadian Heritage Minister Pablo Rodriguez and his department face a significant credibility gap on the file. To be absolutely clear, there is a need for legislation that addresses online harms and ensures that Internet platforms operate in a transparent, responsible manner with the prospect of liability for failure to do so. However, Canadian Heritage has repeatedly fumbled the issue with conduct that raises serious concerns about whether it is fit to lead.
Recent Posts
The Law Bytes Podcast, Episode 231: Sara Bannerman on How Canadian Political Parties Maximize Voter Data Collection and Minimize Privacy Safeguards 
The Law Bytes Podcast, Episode 230: Aengus Bridgman on the 2025 Federal Election, Social Media Platforms, and Misinformation 
The Law Bytes Podcast, Episode 229: My Digital Access Day Keynote – Assessing the Canadian Digital Policy Record 
Queen’s University Trustees Reject Divestment Efforts Emphasizing the Importance of Institutional Neutrality 
The Law Bytes Podcast, Episode 228: Kumanan Wilson on Why Canadian Health Data Requires Stronger Privacy Protection in the Trump Era
