The long awaited government policy direction of the Online Streaming act (formerly Bill C-11) has been released. We provide some analysis.
The Canadian government had been keeping their policy direction out of the public for a very long time. So, the other day, it was quite the news flash that the government was set to release that policy direction. The hope was that there will finally be some clarity on where the government intended on taking things.
Unfortunately, the clarity never really arrived. Instead of a final policy direction, Canadians merely got a “draft” policy direction. Policy directions, by their very nature, can change as the governments objectives change. So, it was always a nebulous moving target. The problem is that this is also a draft policy direction, meaning that it isn’t even anything set in stone. The government, at a later time, can well and truly provide a completely different policy direction at a later time.
To compound the problems with all of this, Canadian regulator, the CRTC, got out ahead of the policy direction and kicked off their series of consultations on the freshly minted Online Streaming Act. The first of the consultations were set to take place next week, leaving participants not only with very little time to put forth a comprehensive framework of ideas, but also left those participants shooting around in the dark thanks to the policy direction that has been missing up until now. Unsurprisingly, participants in the up and coming consultations have asked for extensions citing the tight deadlines and lack of policy direction. Ideally, the final policy direction would be put forward before the hearings took place, but for some, the process looks to be one meant to sew as much confusion as possible.
Still, a not final document is probably better than no document at all – a low bar to be sure. So, what does the draft policy direction say? A number of sources link to this page which… doesn’t really provide much information. However, other sources have been able to provide more interesting links:
https://twitter.com/fanCRTCProfling/status/1666847343717523457
Since last night, a number of pages were adjusted and it is possible to browse through to at least one of those links. What’s more, this link has also changed since last night, providing even more text, though the original text is still offered (albeit with slightly different formatting). It ultimately highlights just how much the situation is evolving. Still, we can happily provide some thoughts on what we read.
User Generated Content
Obviously, the core of the problem with the Online Streaming Act is the regulation of user generated content. The bill itself has the exception saying that user generated content is out, but then adds an exception to the exception, saying that, just kidding, user generated content is in. The policy direction pulls this exact same stunt, but in a different way.
First, here’s this snippet that the government will no doubt tout as proof that user generated content is out:
10 The Commission is directed not to impose regulatory requirements on
(a) online undertakings in respect of the programs of social media creators, including podcasts; and
(b) broadcasting undertakings in respect of the transmission of video games.
Now, you might look at this and say that user generated content is finally out. The problem is, like so many other aspects of the process, the text is designed to mislead. What do we mean? Well, when you pair that passage with the following passage, things suddenly get more confusing:
6 The Commission is directed to consider both established and emerging means of discoverability and showcasing to promote a wide range of Canadian programming. In making regulations or imposing conditions in respect of discoverability and showcasing requirements, the Commission is directed to prioritize outcome-based regulations and conditions that minimize the need for broadcasting undertakings to make changes to their computer algorithms that impact the presentation of programs.
This passage may seem innocuous at first, but it actually directly contradicts the above passage. How is that so? As most users of social media platforms know, the algorithms automatically generate recommendations for a user, showcasing content that the system thinks will be of interest to them. The problem is that the system can’t necessarily promote every piece of content it thinks is relevant to the users taste in content. This would overwhelm the user. So, instead, it recommends a handful of pieces of content, like, say, 8 video’s instead. What’s more, if a user is watching one video in particular, the platform can offer other video suggestions to keep the user interested. This happens on platforms like YouTube, for instance.
This has a big impact on how users discover content. Creators want their video’s to pop up in recommendations because the more people who see their video’s as recommended content, the more likely that users are going to click on that video. The problem arises is if there is an intervention by the government on such a system. So, instead of 8 chances that your video will pop up as a recommendation, the Canadian government might impose a quota of government certified “Canadian Content” or “CanCon” to show up. Let’s say that number is 30%. That means two or three video slots will be taken up by government certified speech that must be promoted.
Now, instead of 8 chances, your content will only have 6 or 5 chances of appearing in recommendations. In terms of absolute impressions (number of times users had your thumbnail loaded in the example of YouTube), this will negatively impact that number. What’s more, it decreases the chances you’ll get those same number of clicks rolling in. This is what many mean by the law downranking content.
So, can you have both contradictory passages remain true? Yes. By artificially boosting government certified speech, the government is fulfilling the mandate of “showcasing to promote a wide range of Canadian programming”. At the same time, there is no additional regulation being imposed on user generated content because nothing new is being applied to user generated content. The net effect remains the same. Government certified speech is mandated to be showcased at the top of the rankings while user generated content gets pushed to the back of the line.
While the negative impacts will be most felt for creators, users will also be hugely negatively impacted by this. If you, as a user, are looking up how to video’s for, say, open source 3D program Blender, under normal conditions, you’ll probably find the user Blenderguru who does, in fact, provide excellent tutorials for Blender. Here’s the problem: Blenderguru is from Australia and won’t likely be considered “Cancon” by the government.
So, instead, the search algorithm is going to have to pick from a much smaller pool of content that is government certified speech and present those as top search results instead. Will you get a video tutorial about Blender? Who knows? Maybe you’ll get something that has some vague resemblance to the search results or get a completely random result. As a result, your search results may not show the best the platform has to offer thanks to such a government mandate. You may be forced to scroll through several results just to find what they were originally looking for. You may be fine with a bunch of scrolling, but others are going to give up and assume that the search feature is broken, degrading the user experience.
Additionally, if you are subscribed to a bunch of poker channels like Daniel Negreanu or Doug Polk, the front page may suddenly have poker video’s from different creators or video’s that don’t pertain to Poker at all. The front page will feel less personalized and more like a place for spam. After all, are you really going to Youtube every day to watch Canadian Heritage Minutes because the government said so or are you, in this scenario, wanting to watch some poker videos? Probably the latter.
Ultimately, this puts us back to square one of the debate all over again.
Algorithms
Another major point of contention, and this relates to the first point, is the issue of algorithms in general. This was covered by section 6 which I quoted above, but I’ll quote again here for convenience sake with some emphasis added:
6 The Commission is directed to consider both established and emerging means of discoverability and showcasing to promote a wide range of Canadian programming. In making regulations or imposing conditions in respect of discoverability and showcasing requirements, the Commission is directed to prioritize outcome-based regulations and conditions that minimize the need for broadcasting undertakings to make changes to their computer algorithms that impact the presentation of programs.
This is the wording that absolutely kills supporters of the Act and the government. Over a year ago, there was a debate on whether or not the CRTC is meddling with the algorithms or not. Ultimately, the truth in the matter is the CRTC has every intention of getting their dirty hands on the algorithms, but they don’t want to come out and say that. So, the law was crafted to say that the CRTC can’t mandate the use of a specific algorithm. Instead, they are mandating the outcomes.
In practice, the difference between demanding the use of a specific algorithm and demanding a specific outcome is non-existent. All you are really doing is differentiating who is going to be the person that does the dirty work. In this case, the platforms have to manipulate the algorithms to fulfill whatever goal the CRTC has set out as an order. In fact, nearly a whole year ago, the then CRTC Chair and spelled out exactly this very scenario, effectively admitting that critics have been correct all along.
Some senators, at the time, were annoyed and suggested that the chair inadvertently muddied the waters, trying to deflect this debate to the classic “he didn’t mean what he said” scenario. However, critics have been saying all along that the CRTC was, for once, being truthful on the debate and this was the intention the entire time. Now, the policy direction basically spells out, yet again, that the CRTC is to demand specific outcomes of the algorithms in question. If a Justin Trudeau campaign ad is deemed priority speech, then the CRTC can mandate that no matter what happens within the machinery of the platforms, those ads must be put at the top of the list (yes, an extreme example, but not an example that is impossible, either.)
So, once again, the critics were vindicated here on this angle. The government is expecting the CRTC to meddle in the outcomes of the algorithms. This despite the insistence that this is not the case.
Social Media Creators
Now, a lot about the bill was about pulling money out of the platforms and handing that money over to traditional producers of content. We’re talking about old television stations, traditional over the air radio, or members of the old guard in the production of other kinds of content. The system was designed to effectively take money from the very creators that generate the value of platforms and redirect it to those that provide little to no value to those platforms. This, of course, is controversial for obvious reasons as we are getting into the government picking winners and losers – specifically in this case, who gets the financial rewards for content produced on social media?
The draft policy direction really doesn’t move the needle on this front. It essentially pays a little bit of lip service to creators. Here’s the relevant sections I was able to find:
social media creator
means a person who creates programs that are primarily intended for online distribution as user-uploaded programs through social media services. (créateur pour les médias sociaux)
12 In exercising its powers under section 11.1 of the Act, the Commission is directed to
[…]
(j) support activities and services — including training and development activities, conferences, the activities of organizations that represent creators and the development of digital and open-source tools and solutions — that support Canadian creators of audio or audio-visual programs for broadcasting by broadcasting undertakings, including social media creators.
This is a bit more confusing because Section 11.1 of the Online Streaming Act only deals with traditional broadcasters. One thing, however, is clear: if a digital first creator goes asking around for money under the system – if only a few hundred dollars to pay for software licensing and equipment – they’ll be out of luck. However, the government is willing to find some training materials for those creators – a service that is generally already supplied by not only the platforms, but also other creators in the first place. It’s difficult to really know if such services provided the government would ever actually change much on this front. If anything, this stuff is more directed at traditional producers who may or may not take an interest in the activities of producing content for social media afterwards.
At the end of the day, the actual concrete help provided is very much tilted towards the traditional producers of content. It siphons off money from the platforms and directs it all away from the creators who are actually on there to the creators who have little to nothing to do with the success of those platforms. In fact, some of those traditional producers spend a fair bit of time railing against the very platforms they are taking financial aid from in the first place. It’s actually kind of a perverse situation: the platforms are forced to pay producers who badmouth them all day long.
Concluding Thoughts
If the government was actually hoping to move the needle with this, that never panned out. If anything, this document ultimately resets many of the biggest debates back to square one. Who benefits? The legacy producers. Who suffers? The users and digital first creators. Just to pour salt on the wound, the document isn’t even final, meaning that it can very easily change to say something different. For all we know, it could get better (not likely) or maybe even worse (where things are trending). On top of that, this document, even if final, can change again if the government decides to change its objectives on top of it all. So, this will always be a moving target.
Either way, this document will do nothing to address the fears brought up by many interested parties. At least we have a year or two to enjoy social media before things hit the fan.
Drew Wilson on Twitter: @icecube85 and Facebook.
