AIDA is a political process error. Committee members should refuse it.
This is the fifth in a series on AIDA. Posts one, two, three, and four are here. And in related writing, a recent piece on AIDA in CCPA’s The Monitor.
AIDA goes to INDU Committee
Next stop for AIDA is INDU Committee, it will likely arrive there in the fall. What this means is that the next set of decision makers on this file are the 12 INDU Committee members. They are as follows:
Joël Lightbound, Liberal (Chair), Rick Perkins, Conservative (Vice-Chair), Sébastien Lemire, Bloc Québécois (Vice-Chair), Nathaniel Erskine-Smith, Liberal, Andy Fillmore, Liberal, Iqwinder Gaheer, Liberal, Bernard Généraux, Conservative, Viviane Lapointe, Liberal, Brian Masse, NDP, Tony Van Bynen, Liberal, Brad Vis, Conservative, Ryan Williams, Conservative.
(There is also a long list of associate members of INDU Committee. I don’t know what they do or what that means so let’s park that for today and roll with the notion that these 12 members are of primary importance re: AIDA and if anyone has more on associate member roles I’m here for it)
Before we move on, a note to keep in mind is that there was an effort by some — and still exists as a belief by some — that the consideration of AIDA should have gone to the ETHI committee. ETHI committee has a broader remit than industrial policy. Unfortunately that did not happen. So we, as a country, are looking at AIDA strictly through an industrial policy lens.
AIDA was drafted and currently lives in the wrong place — ISED
As the conversations about AI have swirled around at high speed, and in the last six months especially, it would not be surprising to me at all if ISED itself is deeply uncomfortable with AIDA. The Ministry has likely become increasingly aware that they are being asked to hold a bag with this bill — in terms of societal consequences — that they have no business holding.
ISED is, however, politically badly positioned to kill AIDA. I don’t even know it they’d be allowed to walk it back at this point. But INDU Committee could do them a big favour by shutting down AIDA so everyone can stop and think about where to best locate, and start, the proper process for AI conversations. The conversations necessary to have prior to any regulation drafting.
I am curious about whether there are any parliamentary ethics touch points that would help INDU Committee members see that they are not tooled properly for what they are being asked to sign off on. If you have thoughts on this please get in touch.
Process History on AIDA is Too Light
One of the hallmarks of good public process and institutional trust building is a coherent narrative about how and why something like a new policy or law has been created. It ties past efforts and activities to current and future work. In the case of AIDA, there has been little effort from ISED to create that clarity, other than: waving generally in the direction of the Digital Charter conversations, waving even more broadly at the opportunities for the Canadian AI industry, then also waving at the specific work that the Advisory Council of AI was mandated to do in terms of public outreach.
No part of these stories have the depth required for adequacy to do something as consequential as provide the rationale for a new law. What was it that triggered the act of glomming AIDA onto C-27? Who wrote it? Answers to these questions will help provide some insight into ISED’s rationale. These answers may also help the committee get a handle on the many angles of this conversation that were not (at least not in any way the public can see) included in the drafting process.
If committee members are willing to acknowledge that this topic cannot be viewed nor addressed appropriately from within the industrial expansion frame, ethics in practice would look like INDU Committee refusing ISED’s proposal as drafter. Committee staff have a chance to move the practice of ethics into a more positive rather than punitive direction — to include contextual and procedural accountability in parliament. Committee staff for each of the MPs on this committee do have individual power and responsibility in terms of assessing and understanding the broad context at play for this topic.
Public, Civil Society, and Expert Participation at INDU Committee
Process wise, here’s what happens next. Sometime in next few months (?) the INDU committee will meet on C-27, including AIDA. As of now, the committee is receiving briefs from the public to collect input on the topic. 8 briefs have been submitted so far.
The analysts that work as INDU committee staff will do research — which includes reviewing these briefs — then write a report for the committee about the proposed law. The proper title of the people writing that report is library of parliament analysts. I believe it’s two of them writing the report. Anyone can submit a brief to be considered in their report. To do this, email your brief to the clerk of the committee: INDU (AT) parl.gc.ca
Briefs are often submitted on behalf of organizations, but can also be submitted by individuals. Whether you submit as an individual or as an organization it will be a public document. If you would like to speak to the committee at their meetings about AIDA, you can request to do that through the clerk (same email as above). And if you’re going to submit a brief *and* want to speak, you can submit your brief and make that request to speak in the same email.
Here is the Canadian House of Commons guide for submitting briefs to House of Commons Committees. From the guide: “Committees often receive a large number of briefs when conducting a study. Therefore, the early submission of briefs to the committee is recommended to allow committee members sufficient time to consider them.”
For the whole of Bill C-27, there are currently…8 briefs
One of them is the Office of the Privacy Commissioner of Canada’s Brief on C-27: Submission of the Office of the Privacy Commissioner of Canada on Bill C-27, the Digital Charter Implementation Act, 2022
Their brief is both light on, and supportive of, AIDA. This, despite being one of the few institutional actors that work on behalf of the general public with the capacity to get into the topic of AI. Setting this lack of depth of comment from the office aside, I am fascinated by the fact that the Office of the Privacy Commissioner of Canada has zero higher order qualms about a) the lack of public engagement on AIDA and b) its housing in/drafting by ISED. Insiders would laugh at the idea this office would be critical of the government and this is another flag about how not arms-length our supposedly arms-length bodies are.
Privacy and data protection are only two of many perspectives that can be helpful in general-purpose technology conversations, so this is not to say that office should be leading on this either.
There many many pre-existing regulatory regimes in effect in Canada. Some of them might be able to be expanded to address some of the issues raised by the use of AI. ISED mentions three of them in its companion document (Health Canada, Office of the Superintendent of Financial Services, Human Rights Commissions). But it’s like two sentences of waving at them with footnotes, and them a jump to the idea of regulatory gaps.
Why aren’t we seeing how existing regulatory actors were engaged in the drafting of AIDA? Were they? Why don’t they have briefs in on the topic? If the answer to this question is that there was a lot of internal consultation done across existing ministries then great, let’s see the results of that work both explained and published somewhere. There is no showing of the work going on here. It’s just hand waving at generalizations and assertion of authority and necessity. This is not adequacy for a new law.
To bring this back to INDU committee: when there is an institutional oversight adequacy/capacity problem, and major process problems, the responsible thing to do is step back and revisit the process. Instead, the people involved in this effort are doing that thing where everyone that should have alarm bells going off because of the problematic closed circuit of the entire AIDA undertaking is pretending all is fine. The current approach and likely response to the issues being raised about AIDA’s lack of depth is that we’ll deal with it all after the law is passed.
That this approach — we’ll do it later — with A LAW — is defensible by our government shows an uncomfortable level of flippancy about public process and law-making.
In Conclusion and Background Reading on Process F
AIDA was written to be, functionally, a press release. it’s a real mind bender to have to take it seriously, but here we are.
AIDA is a specifically shaped problem for our country. It’s status quo preserving. Which is awful. On top of that is the potential missed opportunity to intervene/shape how we deal with AI *way* more thoughtfully.
There is nothing contentious — nothing, zero — about advocating for this process to be restarted. Particularly given the currently proposed timeline ISED has outlined. Nothing.
Background Reading
Some months ago, ISED published this document to try to explain what they are doing with AIDA, and why: The Artificial Intelligence and Data Act (AIDA) — Companion document
ISED’s Advisory Council on Artificial Intelligence
From that Advisory Council — Learning Together for Responsible Artificial Intelligence — Report of the Public Awareness Working Group (Feb 2023)
These three documents give you a decent part of ISED’s story about the background of AIDA.