With the introduction of age verification, I ran an analysis of the bill. Michael Geist did the same and noted the age estimation provisions.
Yesterday, I not only reported on the surprisingly early re-introduction of Canada’s age verification bill, but also ran an analysis of its provisions. In short, the bill is an absolute hot mess by, among other things, envisioning technology that doesn’t exist, putting independent websites on the hook for increased operating cost, having no real enforceable privacy safeguards, and being completely unconstitutional with the removal of legally protected speech. Unsurprisingly, I concluded that it’s a bad bill.
One of the things I like to do after a big analysis is to look around to what others are saying about the bill as well to see if they raised some interesting points as well (not to mention to see how often my analysis agrees with their analysis). After all, bills are very often complex and any one person can easily miss something. That’s why it’s great to have multiple eyes on such a bill. Today, I found out that Michael Geist also ran his analysis and it was an interesting read.
The one point that I did miss that Geist caught was the language surrounding age estimation (though it was captured in my quotes):
Note that Bill S-209 has expanded the scope of available technologies for implementation: while S-210 only included age verification, S-209 adds age estimation technologies. Age estimation may benefit from limiting the amount of data that needs to be collected from an individual, but it also suffers from inaccuracies. For example, using estimation to distinguish between a 17 and 18 year old is difficult for both humans and computers, yet the law depends upon it. Given the standard for highly effective technologies, age estimation technologies may not receive government approvals, leaving only age verification in place.
I’m personally back and forth on these comments. On the one hand, Geist is absolutely correct that age estimation technology (namely ones that use AI) are generally not accurate. There’s plenty of research surrounding that. One study, which was published in Nature, showed that such technology had issued guessing people’s ages:
Our estimates of a person’s age from their facial appearance suffer from several well-known biases and inaccuracies. Typically, for example, we tend to overestimate the age of smiling faces compared to those with a neutral expression, and the accuracy of our estimates decreases for older faces. The growing interest in age estimation using artificial intelligence (AI) technology raises the question of how AI compares to human performance and whether it suffers from the same biases. Here, we compared human performance with the performance of a large sample of the most prominent AI technology available today. The results showed that AI is even less accurate and more biased than human observers when judging a person’s age—even though the overall pattern of errors and biases is similar. Thus, AI overestimated the age of smiling faces even more than human observers did. In addition, AI showed a sharper decrease in accuracy for faces of older adults compared to faces of younger age groups, for smiling compared to neutral faces, and for female compared to male faces. These results suggest that our estimates of age from faces are largely driven by particular visual cues, rather than high-level preconceptions.
On the other hand, backers of this legislation are also known to continually lower the standards of how good a technology needs to be in order to meet the threshold of “highly accurate”. Indeed, throughout the age verification debate during the last government, supporters argued that age verification technology that would suit the bills interest do exist today and that the bill is just a reasonable step forward in protecting youth from the evil demonic mind combusting imagery of showing some skin. This obviously ran contrary to the plethora of evidence that showed that the technology meeting the thresholds set out in the bill doesn’t really exist. This is because it either fails on the privacy front, the effectiveness front, or both. France’s privacy regulator, CNIL, found that out a while back.
So, in other words, supporters and lawmakers pushing this bill were simply trying to pretend that such technology does exist and is willing to settle for just any old “solution” even if it is technically a poor choice. So, even if age estimation tools aren’t up to the task, supporters will simply slap a “good enough” stamp on it and pretend that it fits the need (probably in part because they feel that it’s not their problem to deal with afterwards). So, I think that is a huge problem in the decision making process.
Generally speaking, though, it seems Geist is on the same page as what I found in the bill. It’s an unconstitutional mess filled with privacy and expression concerns.
One thing I will note today, is the point that the government really shouldn’t be in the business of deciding what is and is not appropriate viewing material for people when it comes to legally protected speech. Pornography is legally protected speech (with obvious exceptions of course). It is up to parents to govern what their children can and cannot see.
On an individual level, there are technologies available today that fit that bill. There’s Net Nanny, Qustodio, Bark, and many other’s out there. If you are a parent that is worried about what your kid sees online, there are technological solutions out there. You can limit the different websites your kid can browse, you can set time limits on routers, install third party solutions, and so on and so forth. Parents aren’t helpless here (despite what the bills supporters might argue). The only limitation is what the parents will bother doing.
Where things run start running afoul of the Canadian Charter is when the government begins dictating what people can and cannot see. While the language in the bill keeps trying to say that blocking orders are to prevent minors from accessing certain sites, the only practical solution is if it is blocked from viewing by all people, both of legal age and not. That’s a big part of why there is the concern of constitutional protections for free speech. What’s more, it’s the adults that pay the price of their own privacy. Either way, the government has no place in dictating what legally protected speech people can and cannot view. That is precisely what drives the criticism of this bill.
At any rate, it’s definitely an interesting perspective. It adds to a growing consensus on how bad this bill is and it is a different set of eyes on the same issue. Definitely a valuable contribution to the debate.
