TL;DR

The UK Labour Party has been running an ersatz signals intelligence (SIGINT) operation to identify and exclude members and supporters that they don’t want voting in their leadership election; people who under some sort of criteria are identified as enemies of the party. This should be terrifying, as the difference between enemy of the party and enemy of the state is an election away (and if that party gets its hands on the levers of state power then it levels up to a full capability SIGINT operation).

I don’t think it’s overstating things to say that this is the most significant post Snowden example of the impact of mass surveillance on democracy.

Background

A vote of no confidence in leader Jeremy Corbyn from the members of parliament (MPs) in the Labour Party has triggered a leadership election. I won’t go into the details here, as it’s covered in its own Wikipedia entry. The party has been crawling through social media to identify unwanted members and supporters and exclude them from the party or leadership election process. The Twitter hashtag for this process is #LabourPurge2.

I applied to be a ‘supporter’ of the Labour Party as I feel it’s important that the country has an effective political opposition and I wanted to have a say in the process. Here’s the letter I got:

I’ll use another post to pick apart my 5 Jul 2016 Tweets for what might be considered ‘racist, abusive or foul language’ (and whether any of that could be construed as being aimed at Labour Party members) as I don’t want to get sidetracked on this post.

The point here is that they went a few weeks back before finding something that matched the criteria to exclude me. The interesting questions I’ll spend the rest of the post on are: who are ‘they’, what are ‘the criteria’, and how is this all being done?

‘They’ are watching, in secret

Somewhere at the core of the Labour Party there’s a bunch of people who decided that this was necessary, and somewhere else there’s a bunch of people actually doing the spade work. I’m not going to waste time here speculating about the motivations of the former group. The interesting part is that even with open source collection (like public tweets or Facebook posts) and sophisticated analysis tools there needs to be some degree of human assessment. The humans doing that are (like in the case of grown up SIGINT operations) having to do their work in secret, because the reveal the details would be political dynamite. The last thing that the group directing this will want is their criteria being spilled out (like say the ATOS fit-to-work checks), so they can’t just outsource it to the lowest bidder – a high quality circle of trust needs to be established here.

This is where I necessarily stray into speculation (because the facts are hidden):

1. I expect that the work is being done by a law firm, because they provide clerical/analytical skills for hire, and understand (and most importantly respect) non disclosure agreements. Another possibility would be a ‘big 4’ type consulting firm.
2. The £25 paid by ‘supporters’ is being used to pay the professional services fees associated with this whole exercise.
3. Somebody had already figured out the cost of this exercise hence the steep jump in the ‘supporter’ fee (it was £3 last time).

What are the criteria?

The precise criteria of what makes an ‘enemy of the party’ are secret, and that’s a big part of the problem, but we can get some feel for it by observing the responses to and reactions of those being excluded.

Isn’t this just about enforcing the rules?

I’ll take a brief diversion here on the topic of party rules and how they’re interpreted. One might argue that the criteria are publicly and transparently embodied in the party’s rules, and this whole exercise is simply about ensuring that the rules are properly enforced. It’s quite fair and proper that a party doesn’t want people who are racist, abusive or use foul language; but each of those things is open to some degree of interpretation, and it’s in that interpretation that the true criteria for this exercise lie. When I pick through my tweets for 5 Jul the only thing I can be sure of is that I didn’t press the racist button – the criteria are subtle enough that it’s not clear to me which tweet got me found out as an enemy of the party. What I am pretty sure of is that if it hadn’t been 5 Jul then there would have been something incriminating further back in my timeline.

It’s harder on members

I was only trying to be a ‘supporter’ before being found out as an enemy of the party and excluded. People who’ve switched allegiances from other parties to Labour have found themselves excluded for their past support of those previous parties. Here’s an example from Ben Crawford (picked from the top of the pile on #LabourPurge2):

Here lies the point about the dangers of using SIGINT like this – people are being found guilty for their sins of the past. It creates a world where nobody can engage in politics and change their mind (because they think something changed or they were presented with a better argument, reasoning or set of data). It means that only the purist ideologues can engage in the political process. It is exclusionary.

Conclusion

I’ve used the word ‘exclude’ a lot in this post, because that’s what this is about – exclusion. Rules against people who are racist, abusive or use foul language are there to create an inclusive environment, but they’ve been twisted into a set of SIGINT filters to identify ideological enemies and exclude them. We’re getting a front row seat here to what happens when modern SIGINT is used for political purposes, and I’m writing this because I don’t like its impact on democracy.