Who benefits from “democratising technology and innovation”? Answer: it’s complicated.

In May 2019 (1) DiEM25 published a Green Paper titled “Technological Sovereignty: Democratising Technology and Innovation”. The purpose of the paper is to propose the policies and visions of DiEM25 for “a democratic future of technologies — and of the future of democracy in a technologised society”.

The three examples of how better the resulting world would feel like are, from left to right:

• “Paul 33, Berlin: freelance writer and part-time chef de cuisine”
• “Pierre, 41, Paris: scientist with a PhD in biological engineering”
• “Diane, 29, Estonia: junior developer with a considerable skill set”

These three beneficiaries of that DiEM25 vision are all [relatively] young urban professionals; all just born with hard to find, hard to automate, highly valuable skills; all highly educated; all, consequently, quite more likely than others to already have a good income and quality of life. Also, none of their life descriptions makes any mention of partners, children or elder relatives to care for.

Those three people made me immediately remind what, back in 2012, I called the “Zeroeth Reason why the Green Party won in Germany”: they won, I believe, simply because Germany was “in much better economic shape than most other countries these days” (and so it could afford the luxury to vote Green). I am afraid that, as presented in that paper, this is a similar case: where are families, senior or disabled citizens, victims of technological unemployment, working poors and so on?

As is, the roadmap in that paper seems to “democratise technology and innovation” only for privileged people who were already having a good life anyway. I am sure that this is not the intention or vision of the authors. But I’m afraid this is how the vision will be perceived by too many people, all around Europe and beyond, who for this reason will not understand why they should ever vote to “democratise technology and innovation”.

On a related note…

Section 2.2.3 of the paper introduces a new framework for digital citizens' rights. Those rights include, of course, a Right of Equal Treatment that states, among other things, that:

“If algorithmic services provide outputs of consistently lower value or quality to or about users coming from historically marginalised backgrounds, this constitutes discrimination."

In general, I am slightly skeptical about creating new, digital rights, (or a Right to the Internet, as proposed in Italy). It seems to me an unnecessary complication. The right of any group of people to not be discriminated because of what they ARE already exist, at least on paper, in most “advanced” states.

Why create a parallel right, instead of, for example, update and extend the definition of discrimination to “receiving lower quality services, or being abused, also through or from algorithms”? For the same reason, I don’t really like formulas like “digital citizens”. Ditto for speaking of “users”, or “consumers”, in any context like this. There are citizens, or even better: people, human beings. And they all have equal rights in all spheres of life, period.

This said, and playing devil’s advocate of course, here is my main objection to the quote above: what if an algorithm consistently provided lower value to people coming from historically privileged backgrounds?

What if an algorithm made, say, affluent, white, middle-aged males living in EU capitals systematically pay more for an hotel room, or concert ticket? According to the letter of that sentence, that would not be discrimination, would it now?

Why not just replace “users coming from historically marginalised backgrounds” with something like “to ANY specific group of people”? This is an age (1) in which “Conservatives Allege Big Tech Is Muzzling Them”, why make things even more complicated?

1. I wrote this post on June 3rd, 2019, and this is why I kept that as “publication date”. For technical and personal reasons, however, I could not put it online before July 30th, 2019, when I added the link to the Atlantic article. This is why you see a post “published” on June 3rd that links to something published almost two months later.