Any Privacy Law Is Going To Require Some Compromise: Is APRA The Right Set Of Tradeoffs?
Privacy issues have been at the root cause of so many concerns about the internet, but so many attempts to regulate privacy have been a total mess. There's now a more thoughtful attempt to regulate privacy in the US that is (perhaps surprisingly!) not terrible.
For a while now, we've talked about how many of the claims from politicians and the media about the supposed (and often exaggerated, but not wholly fictitious) concerns about the internet are really the kinds of concerns that could be dealt with by a comprehensive privacy bill that actually did the right things.
Concerns about TikTok, questionably targeted advertising, the sketchy selling of your driving records, and more... are really all issues related to data privacy. It's something we've talked about for a while, but most efforts have been a mess, even as the issue has become more and more important.
Part of the problem is that we're bad at regulating privacy because most people don't understand privacy. I've said this multiple times in the past, but the instincts of many is that privacy should be regulated as if our data were our property." But that only leads to bad results. When we treat data as property, we create new, artificial, property rights laws, a la copyright. And if you're reading Techdirt, you should already understand what kind of awful mess that can create.
Artificial property rights are a problematic approach to just about anything, and (most seriously) frequently interfere with free speech rights and create all sorts of downstream problems. We've already seen this in the EU with the GDPR, which has many good characteristics, but also has created some real speech problems, while also making sure that only the biggest companies can exist, which isn't a result anyone should want.
Over the last few weeks, there's been a fair bit of buzz about APRA, the American Privacy Rights Act. It was created after long, bipartisan, bicameral negotiations between two elected officials with very different views on privacy regulation: Senator Maria Cantwell and Rep. Cathy McMorris Rodgers. The two had fought in the past on approaches to privacy laws, yet they were able to come to an agreement on this one.
The bill is massive, which is part of the reason why we've been slow to write about it. I wanted to be able to read the whole thing and understand some of the nuances (and also to explore a lot of the commentary on it). If you want a shorter summary, the best, most comprehensive I've seen came from Perla Khattar at Tech Policy Press, who broke down the key parts of the bill.
The key parts of the bill are that it takes a data minimization" approach. Covered companies need to make sure that the data they're collecting is necessary" and proportionate" to what the service is providing. This means organizations making over $40 million a year, processing data on over 200,000 consumers, and that transfer covered data to third parties. If it's determined that companies are collecting and/or sharing too much, they could face serious penalties.
Very big social media companies, dubbed high impact social media companies," that have over $3 billion in global revenue and $300 million or more global monthly active users, have additional rules.
I also greatly appreciate that the law explicitly calls out data brokers (often left out of other privacy bills, even though data brokers are often the real privacy problem) and requires them to take clear steps to be more transparent to users. The law also requires data minimization for those brokers, while prohibiting certain egregious activities.
I always have some concerns about laws that have size thresholds. It creates the risk of game playing and weird incentives. But of most bills in this area that I've seen, the thresholds in this one seem... mostly okay? Often the thresholds seem ridiculously low, covering small companies too readily in a way that would create massive compliance costs too early, or only target the very largest companies. This bill takes a more middle ground approach.
There are also a bunch of rules to make sure companies are doing more to protect data security, following best practices that are reasonable based on the size of the company. I'm always a little hesitant on things like that because whether or not a company took reasonable steps is often viewed through the lens of retrospect, after some awful breach occurs, when we realize how poorly someone actually secured their data, even if upfront it appeared secure. How this plays out in practice will matter.
The law is not perfect, but I'm actually coming around to the belief that it may be the best we're going to get and has many good provisions. I know that many activist groups, including those I normally agree with, don't like the bill for specific reasons, but I'm going to disagree with them on those reasons. We can look at EFF's opposition as a representative example.
EFF is concerned that it does not like the state pre-emption provisions, and also wishes that the private right of action (allowing individuals to sue) would be stronger. I actually disagree on both points, though I think it's important to explain why. These were two big sticking points over previous bills, but I think they were sticking points for a very good reason.
On state pre-emption: many people (and states!) want to be able to pass stricter privacy laws, and many activists support that. However, I think the only way a comprehensive federal privacy bill makes sense is if it pre-empts state privacy laws. Otherwise, companies have to comply with 50+ different state privacy laws, some of which are going to be (or already are) absolutely nutty. This would, again, play right into the hands of the biggest companies, that can afford to craft different policies for different states, or that can figure out ways to craft policies that comply with every state. But it would be deathly for many smaller companies.
Expecting state politicians to get this right is a big ask, given just how messed up attempts to regulate privacy have been over the last few years. Hell, just look at California, where we basically let some super rich dude with no experience in privacy law force the state into writing a truly ridiculously messed up privacy law (then make it worse before anything was even tested) and finally... give that same rich dude control over the enforcement of the law. That's... not good.
It seems like the only workable way to do this without doing real harm to smaller companies is to have the federal government step in and say here is the standard across the board." I have seen some state officials upset about this, but the law still leaves the states' enforcement powers on the more national standard.
That said, I'm still a bit wary about state enforcement. State AGs (in a bipartisan manner) have quite a history of doing enforcement actions for political purposes more than any legitimate reason. I do fear APRA giving state AGs another weapon to use disproportionately against organizations they simply dislike or have political disagreements with. We've seen it happen in other contexts, and we should be wary of it here.
As for the private right of action, again, I understand where folks like the EFF would like to see a broader private right of action. But we also know how this tends to work out in practice. Because of the ways in which attempts to stifle speech can be twisted and presented as privacy rights" claims, we should be wary about handing too broad a tool for people to use, as we'll start to see all sorts of vexatious lawsuits, claiming privacy rights, when they're really an attempt to suppress information, or to simply attack companies someone doesn't like.
I think APRA sets an appropriate balance in that it doesn't do away with the private right of action entirely, but does limit how broadly it can be used. Specifically, it limits which parts of the law are covered by the private right of action in a manner that hopefully would avoid the kind of egregious, vexatious litigation that I've feared under other laws.
Beyond the states and the private right of action, the bill also sets up the FTC to be able to enforce the law, which will piss off some, but is probably better than just allowing states and private actors to be the enforcers.
I do have some concerns about some of the definitions in the bill being a bit vague and open to problematic interpretations and abuse on the enforcement side, but hopefully that can be clarified before this becomes law.
In the end, the APRA is certainly not perfect, but it seems like one of the better attempts I've seen to date at a comprehensive federal privacy bill and is at least a productive attempt at getting such a law on the books.
The bill does seem to be on something of a fast track, though there remain some points of contention. But I'm hopeful that, given the starting point of the bill, maybe it can reach a consensus that no one particularly likes, but which actually gets the US to finally level up on basic privacy protections.
Regulating privacy is inherently difficult, as noted. In an ideal world, we wouldn't need regulations because we'd have services where our data is separate from the services we use (as envisioned in the protocols not platforms world) and thus more in our own control. But seeing as we still have plenty of platforms out there, the approach presented in APRA seems like a surprisingly good start.
That said, seeing how this kind of sausage gets made, I recognize that bills like this can switch from acceptable to deeply, deeply problematic overnight with small changes. We'll certainly be watching for that possibility.