There is an ongoing discussion on human rights on the Internet on the IRTF HRPC mailing list that I want to express an opinion on.
I would also like to stress that this is not an official position of the Interpeer Project. Although we are yet small, there exists already a variety of positions amongst contributors on all kinds of topics. No, this is a purely personal opinion.
For context, the Internet Research Task Force (IRTF) is a sibling to the Internet Engineering Task Force (IETF) that does not get a lot of publicity. While IETF concerns itself with standards on interoperability of current networks, IRTF looks further into the future. The HRPC list concerns itself with Human Right Protocol Considerations, i.e. any topic that relates to how protocol design work may impact or support human rights.
Since the Interpeer Project aims to help bring about a human centric Internet, this list is of utmost interest to me. As a consequence, of course I brought myself into the discussion.
The discussion topic is censorship, and ties into current political events. As a result of the Russian Invasion of Ukraine, Ukraine has asked the international community for help. One such request was to shut down (parts of) the Russian Internet as part of their cyber defence.
The discussion was sparked by a paper (linked in the first post) authored by some noteworthy Internet related organizations on how various technical means for sanctioning Russian access to the Internet do or do not help or hinder, well, “things” in general. But it also concludes that some technical means pose an acceptable risk and could be implemented. They boil down to publishing blocklists.
This, of course, is perceived as censorship and overreach by some parties, and so the discussion ensues.
In my personal estimate, it is important to emphasize a few things here that may get lost in the overall debate.
Goals
Any party that has so far contributed to the discussion shares the same goals.
- One goal is to help reduce harm to people, which mostly refers to physical harm.
- Another goal is to preserve freedom of access to the Internet; this is in some ways identical to preserving the Internet as a network of networks.
It’s worth highlighting that there is no disagreement on these goals. I think it’s also of interest that these goals align with the Universal Declaration of Human Rights, in particular Articles 3-5 and 12 (and/or 19) respectively.
This entire discussion, then, is the process of balancing two human rights against each other, where none of them has more value than the other.
Positions
In the discussion, one can also see a few positions crystallizing in opposition of blocklists. I may not be able to provide a complete or fair list below, but some opposing positions include:
- Human rights concerns are the purview of politics, not technology folk.
- Anything technology folk do can be used or abused by politics.
- Politics has a tendency to use technology for political advantage without understanding it.
This leads some folk to the following conclusion: if politics uses technology without understanding, that constitutes abuse of technology folk’s efforts, so in order to preserve in particular the freedom of access to the Internet, it’s best for technology folk no to undertake efforts that can be abused in this way.
In and of itself, that is a fair assessment. But it also leaves the decision space entirely to the people who are not experts on Internet technology. I cannot see how this will in practice help preserve anything about the Internet’s best aspects.
The paper under discussion takes an entirely different position: it starts from the point of view that decisions about the goals of the Internet cannot be left to politics alone, so tech folk must engage with politics. If politics demands some kind of sanctions being enacted, then the best route is to provide politics with the choice of sanctions that do the least damage possible, and actively preserve that which needs to be preserved.
Martial Arts
Either of the above position is a defensive position. In martial arts, depending on the style involved, one will be taught two types of defensive moves when contact is unavoidable, blocking and deflection. In practice, the two are not as clearly separated as one might think, but the categories serve an important purpose for illustration.
When blocking, the force of the attack is effectively absorbed. What makes the block effective at averting damage is that the block occurs at a site of the defender’s choice, where the kinetic energy may more readily be absorbed.
On the other hand, deflection aims at redirecting the force of the attack elsewhere, so that - ideally - no part of the defender’s body needs to absorb any kinetic energy.
Either form of defence is valid in and of itself.
There are, however, reasons why deflection is the superior choice in the long term. One is that any absorption of an attack’s energy eventually exhausts the defender. It’s best to avoid this, though this can also be weathered with sufficient stamina.
The more important reason is that the kinetic energy of the attack does not stop when it is deflected; it has to go somewhere. This either leads the attacker to overextend themselves, or to exert effort in pulling back. In either case, the attacker becomes more vulnerable themselves, if only for the briefest moment.
I contend that when engaging with politics to create sanctions, this is essentially a deflective defence against attacks on freedom of access to the Internet. It makes such attacks easier to withstand by sapping less energy, and makes the attacker more vulnerable.
Not engaging on the grounds that it violates some principles, by contrast is a block: it requires civil society to absorb and re-absorb over and over again the same kinds of attack until the defence is ground to dust.
Process and Principles
The above can only hold true if engaging with politics occurs in just the right way. Finding this form is a thoroughly difficult task to begin with, and to make matters worse, any approach once discovered must necessarily adjust as the political landscape changes.
I strongly suspect that proponents of the blocking kind of defence fear this difficulty. I also think they’re right to do so. But I do not agree that this fear is a good reason not to try anyway.
What precise form this should take is beyond me as an individual. But I can offer some insights into what qualities it should have.
- The principles it must be anchored in should not be attackable by politics. I refer to the UDHR above for a reason; while not every nation on this planet has ratified it, it is nevertheless the least contentious law amongst the community of nations, and attacking it is treated as an attack on all. While this does not rule out such attacks, it ensures backing.
- If we learn anything from safety engineering, quality assurance, etc. it is that humans are fallible and machines stupid. The latter point means that any process surrounding a blocklist cannot be safely automated, but needs human eyes upon them. The first part means that no single human should be in charge of providing those eyes.
- From the above follows that a diversity of opinions - from multiple stakeholders involved in the process - provides the best safety net against abuse of such mechanisms.
- People are prone to forgetting what they did, which means any process for creating blocklists must contain provisions for periodic review. Such review must not default to keeping a sanction in place, but the sanction must be defended anew at every renewal.
In summary, additions to a blocklist must essentially be discussed and voted on, as such are the mechanisms of multi-stakeholder consensus building. But I would add that any implementation of blocklists should also require much the same, or we otherwise invite backdoors for abuse or laziness. In short, it should be made illegal to implement the list without undergoing some kind of review process again.
This does not, of course, prevent any implementor from creating a list with the same items and short-circuiting the process in that fashion. However, this cannot be done with the approval of the blocklist’s originator, and therefore raises the question why a multi-stakeholder process firmly entrenched in human rights considerations was circumvented in the first place.
This deflects the attack back onto the attacker.
Effectiveness
Criticism of blocklists in general are mostly discussed in the beginning, but one specific criticism remains: that of effectiveness. As some commenters wrote, a blocklist is likely going to be either overly broad and damaging, or ineffective. By introducing a slow and considered approach to populating the list as hinted at above, one may achieve a less damaging scope, but simultaneously render the list less effective.
This is true, and somewhat by design.
The criticism rests upon a false dichotomy between the goals mentioned at the outset. While it is true that blocking some particularly bad actors will reduce harm to people, more harm will be averted by not blocking everything else.
That is, the list should not be an effective shield against everything. No life can thrive in a lightless bunker. Instead, it should aim only to avert the worst disasters, and let through light and nourishment of any kind.
High effectiveness in protection is a false goal. High effectiveness in balancing human rights against each other is the true aim.
This may make such a blocklist a somewhat symbolic gesture. This, too, has been raised as a criticism. At which point I have to ask, which symbol has not led to change? The answer is, of course, none of the ones that you can remember.
Conclusion
As I wrote at the outset, none of the above is an official Interpeer Project opinion, it is mine alone. Of course as the initiator of the project, I will undoubtedly wield undue influence over it for the time being. I hope that this will change in time, because I am far from infallible.
I stress this because I do not believe that any part of the Interpeer software stack to come should default to implementing some kind of blocklists, whether arrived at by a well-defined process in deference to human rights or otherwise.
I do believe that in order to create a human centric Internet, such decisions have to move closer to the end-user. It should not be anyone else that decides who to block and why, but every user for themselves.
This point of view makes it all the more important that we can leverage the wisdom of others. A blocklist published by others is the embodiment of such wisdom; they tell us that “here be dragons”. Such a blocklist transparently and carefully curated under strict human rights considerations is almost certainly going to be a better choice than any random blocklist off the Internet (though it should by no means be the only choice).
Which means the creation and publication of it is something of a necessity in the same way that the implementation must ideally be each individual’s choice.
I cannot predict whether the discussion above will lead to any such blocklist, or whether the criteria applied to creating it satisfy my needs, either the ones outlined above or others yet to be discovered.
But I can say with absolute certainty that I would much rather have a disagreeing body of folk all concerned with human rights grudgingly and publicly create one than leave this entirely in the shadows, determined by politics or corporations.
In that sense, I applaud the effort put into the paper that sparked this discussion. It’s the first step in dragging these things kicking and screaming into the light, where they belong.