Discuss macroeconomic impact and equality?

[RByers]

Thank you for all the work that's gone into the TAG ethical web principles, there are lots of points here with which I agree strongly and personally motivate my work trying to improve the web. One area I don't see discussed here is macroeconomic impacts.

Section 2.4 describes how "the web is for all people" but doesn't mention income inequality. The web supports a massive global marketplace for e-commerce and content on which many consumers increasingly rely for both necessities like buying clothing, luxuries like consuming entertainment, and things somewhere in between like access to quality journalism. What position should the TAG take on changes which may impact this economy?

For example, if through changes in the architecture of the web platform, lower income people end up being disadvantaged such as by paying more for their goods and content online while higher income people end up being further advantaged, is that an ethical concern of interest to the TAG? Does the TAG take a position on, for example, the effects which influence how much quality journalism is available on the web freely vs. only behind a paywall?

[RByers]

Looking through open issues I see #68 is similar. @tobie, WDYT is this a dupe I should merge this into yours?

[tobie]

Definitely feels like complementary facets of the same underlying idea. Merging them sounds like the right thing to do. Feel free to edit the title of #68 accordingly if that’s helpful.

[hober]

I think it would be very difficult to write such a principle in a way that doesn't suffer from status quo bias. See also @mnot's thoughts re: "shifting from a principles-based approach to an effects-driven one" in w3ctag/privacy-principles#224 (comment)

[darobin]

I think that @RByers's point is aiming in the right direction, but that there are a few issues at play here.

The first is that §2.4 talks about the web being "for all people" but then goes on to be phrased mostly (though not entirely) in terms of things that the web shouldn't prevent. Put differently, you would want it to be about empowering people but instead it's about not disempowering them. Those aren't the same thing, because people start from different places. It looks like §2.8 would mitigate that but then it's also mostly about mitigating harm.

If we follow the logic of §2.4 to add income inequality, we would end up with something like "People shouldn't need a high level of income in order to access the web" or "the web platform's content must be accessible to people with limited income" or (ew) "we will accommodate people with low income." But these are weak propositions. They say nothing about people actually having agency. If you put all the content in the hands of a big central goop factory and architect a system in which the content that people can access is mostly crap because it's dominated by MFA (Made For Advertising) sites, then you haven't built a liberating system at all — even though it doesn't discriminate based on income inequality.

The second issue is that the macroeconomic framing isn't a great one here. A macro angle would involve summarising ground-level data into major indicators (à la unemployment, GDP) and using that for decision-making. But there's a reason there is broad consensus that something that GDP is a bad metric (and an even worse target): it ignores very important structural details like distribution, it doesn't account for the fact that someone with a health condition might need more income to have agency parity with someone who doesn't, it cannot reflect the fact that some people simply don't get the same opportunities even with income (cf. redlining).

It is possible to support more living that people have reason to value with lower GDP better distributed than with higher more captured. To stick to the same example, it is possible to support more content worth reading/watching/etc. (for people) with significantly lower revenue but a lot less MFA than with higher global revenue that supports greater parasitism.

Third, this is formulated in a biased, very welfare-centric way. For instance, this is given as a hierarchy: "necessities like buying clothing, luxuries like consuming entertainment, and things somewhere in between like access to quality journalism." We have no business setting priorities between these things for people. I'll note that there's a not insignificant body of work in development that places access to quality journalism at the necessity level because it can help increase individual agency, coordination, and hold power to account, all of which have more robust effects in lifting people out of poverty than, say, donating clothing. (Interestingly, none of this thinking requires the content to be free).

These three issues are linked: they all treat people not as the source of action but as mostly passive recipients of our (us powerful people) macroeconomic generosity. Again, I'm not saying that Rick's suggestion that we should think about money is wrong (it's not), but we should be very thoughtful about how we approach this. I have mentioned development a couple times because there's a lot of precedent there of thinking about income inequality, there's also a lot of precedent there for a bunch of powerful technocrats making trade-offs "for all people" that went spectacularly wrong (we've probably all read Seeing Like A State, you don't need the spiel), and that field has also given us an approach that can frame a solution.

I am particularly fond of the capability approach. I won't describe it in great detail, but it focuses on what we might call user agency (some very very early notes on how that translates to the web for the intrepid). Essentially, the focus is on “What each person is able to do and to be?” You are looking for changes that, concretely and on the ground, empower people to do more and to be more, generally the live lives they have more reason to choose. A key focus is on developing substantive freedoms rather than vaporware freedoms (things like you have the right to start a business, but no one will lend money to people like you or you can write whatever you want to write, but a tiny number of companies will decide whether you are read).

To take a concrete example from this discussion (not picking on Rick, it just happens to be there), this has no mention of people or of their freedoms: "Does the TAG take a position on, for example, the effects which influence how much quality journalism is available on the web freely vs. only behind a paywall?" A capability-centric way of reframing that question might be: "Does the TAG take a position on how people can gain better access to more desirable choices of quality content?" (Or something like that.)

Reframing it this way helps us ask the right questions. Sure, whether you have to pay matters, but it tells us nothing about what content people actually get to access. It's not hard to imagine a world in which all content is paywalled but people have greater effective access to higher quality content (thanks to piracy — SciHub works very well) than in a world of free access to an endless morass of SEO-heavy MFA. In other words, free content has the makings of vaporware freedom. If we ask "freely vs a paywall" we will only get answers that are already contained in the question's hidden assumptions. Putting people at the centre opens different questions:

• Is the content they want being produced?
• Do they have access to content that supports their political participation and the exercise of practical reason?
• Do they have agency over how they find that content, over how content is ranked and recommended to them, and is their discovery self-directed?
• Can they easily discuss, share, curate, etc. (the list goes on).

I won't go in to greater detail, this is already waaay too long, but the first part of the point is: when we try to answer these questions from any perspective other than increasing user agency, we risk losing contact with reality (as we all know many arguments about making the Web free as in beer have).

The second part, to return to the original question, I totally agree with Rick's intent (or what I read into it at least) that money should matter and that it should factor into our ethical considerations. We in tech and standards have the bad habit of ignoring money. It turns out, it won't ignore you back. But focusing on a macroeconomic view will just lead us down the path of thinking that we know what we're doing because we have data, which remains a sadly common despite voluminous data showing that panoptic legibility has pretty much always led organisations to make bad choices :) Things I would suggest are both agency-focused, related to money, and in scope for ethical consideration, that might be too specific but could perhaps be surfaced in the doc somehow include:

• Do people gain better access to more desirable choices of better content if it is subsidised? I suspect the answer to that one is "possibly, but it's complicated." The current model uses cross-context data processing to tax high-signal contexts in order to subsidise low-signal ones. It doesn't cover whether that taxation is fair, transparent, accountable, or efficient. Since the subsidy doesn't discriminate on quality (which would be hard and problematic), it doesn't account for how much of the subsidy is lost by legitimate actors of all sizes to MFA. It also affords no control over the intermediaries that administer the taxation (ie. it's literally taxation without representation). I would certainly be very worried if the TAG were to express a position about paywall vs ad-based, there's zero evidence that as a whole it empowers people with better access.
• Are people being paid for their work? People should be compensated fairly for the work they produce. For instance, when a company uses their work, people should be compensated. We could be working on building that into the web today. There's some lightweight work in a CG but it would need to be stronger. Ethically this is pretty fundamental.
• Can people pay for things? Bundling and free access often in fact reduce agency since people have much less of a voice. Functionality gets bundled into apps or suites because we don't have a good solution for revshare nanopayments. This also decreases the value of standards since people are locked inside bundles. Giving people the agency to debundle apps and empowering them to pay for pluggable features would be an ethical step forward.