Neither interoperability nor data sharing will solve transnational monopoly

Every since my awesome if brief visiting professorship at the Tilburg Institute for Law, Technology, and Society in 2019 I've been worrying about why some law professors believe that all data is the same, and that if we forced companies to share theirs that would somehow remedy the inequalities underlying the market dominance of tech giants right now. There are many false premises there – e.g. myth of first-mover advantage (hello IBM, hello myspace, hello Mosaic – yes there are huge network effects, but these can include sudden transitions to new winners.) But the main thing is that it shows an ignorance of the nature of computation. Computation is a physical process requiring, time, space and energy. The tech giants are now enormous global-scale infrastructure, with their own power plants, global fiber-optic networks, satellites, and so forth. 

I agree that interoperability where it is viable can be a partial solution to the problems of antitrust, but "where it is viable" is a very small part of the domain of the excesses and assaults we are seeing against antitrust regulation, even if we only focus on the digital sphere – which to be honest, I think is a mistake. Let me unpack these two thoughts.

There are a few services such as texts and email that you might be able to make fully interoperable. But even if you get to the point of thinking hard about twitter vs linkedin, you might already see the problem with interoperability. These are different platforms with different communicative goals and as such will have different ideal representations. No data system captures the entire universe, nor even the universe of human self expression. 

We should actively want our technical venders only to retain what data is necessary for their mandates (cf. the GDPR.)  And indeed even for all the talk of "the new oil", tech vendors too keep only compressed and filtered versions of all the data they see. This is because retention and transmission take time and energy. The compression and filtering is business-practice specific. Demanding data sharing by e.g. giants to SMEs is really demanding lock-in to particular business models / world views depending on which giant info-ecosystem an SME chooses to deploy within. 

Demanding that all the giants become replicas of each other is actually one thing I've thought of with respect to remedies – make big tech "airbus" each other. But I just don't think this is ecologically sound or pragmatic from a power perspective. You would then just essentially unify them into one bigger monopoly by aligning all their interests. 

The reason I think focussing only on the digital is a mistake is because in my opinion, one of the fundamental problems underlying our present situation is that we are inadequately governing and regulating all kinds of transnational entities. This includes finance, petrochemical, pharma, and consultancies. If we do something that only addresses Google and Facebook, then yes we do address a couple of the very largest bubbles of the moment – though note that interoperability of doc / ppt / xls only helped Microsoft to further power and success.* Yes, it did so via providing us with more consumer choice so I do very strongly approve of that legal decision. The goal of antitrust isn't to destroy the real economic and security value these companies provide. It's to defend further innovation and also the rule of law, hopefully via democracy.
The goal of antitrust isn't to destroy the real economic and security value these companies provide. It's to defend further innovation and also the rule of law, hopefully via democracy.
So actually, there are two problems with focussing too much on a digital-only remedy like interoperability.  First, this lets all the other sectors not being adequately addressed off the hook. I'm suspicious that part of the reason that the market caps / power of big tech have been allowed to get so out-of-control (at least in America) is because the geopolitical power of the peterostates was seen to be surprisingly resurging in the early naughts. There was back then some discussion in the newspapers that perhaps liberal democracy and capitalism couldn't really beat out autocracy coupled with resources and a strong cyber-competent mafia. Now we are hearing about discussions about how separated GAFAM should be from the White House (think for example Schmidt).

But second, Apple, Google, Facebook & Amazon at least benefit most when liberal democracy is strong – when there are many people with enough wealth, education, and freedom to use their products. So these are organisations we should be trying to bring into the fold of understanding and supporting good regulation. My model here is FDR's New Deal and also Bretton Woods. You need enough of the elite on your side to achieve such titanic changes in the regulatory order. This is why I'm personally very worried also about the DMA. We shouldn't be demonising exclusively the digital monopolies. We should be returning antitrust enforcement to its original emphasis on keeping corporations to a size that can be governed by democracies. 

So sure, where viable, make demands about interoperability, and for sure let individuals download complete sets of their own records. But don't play into the libertarian fantasies that there's some magic way that the digital economy could represent information and play nice together that would mean we don't have to remain vigilant about the state of our democracies, or that government has no real utility in an adequately transparent world. Government is a means of coordination, when we deploy it well the rest of our problems become simpler. And government is defined by a monopoly of executive force within a geographic region. There will always be at least a leader if not a complete monopoly in that regard. Let's ensure that we keep strong democratic control over ours.

One of Google's many data centre / renewable power stations (a converted non-renewables plant).
Note the (AFP) link says Amazon is the only of GAFA not showing commitment to renewables.

*note also that no one made or suggested that Microsoft should just have one format that combines xls, ppt, & doc – yet some lawyers seem to think that any startup should be able to utilise the data of any big tech company "as is."


Update 24 Feb 2023: James Bessen has data showing that software makes monopoly more likely in every sector, not just big tech. See his book "The New Goliaths" or this 30 minute video of him talking about the data.

Comments

Anonymous said…
I agree that interoperability isn't a silver bullet: it is not applicable to all types of companies and it is not enough on its own to tame disproportionate corporate power. Yet I think that Strong interoperability is a very potent instrument that is underestimated and underused by regulators.

Facebook claims that their set of APIs are proof of interoperability and allows a third-party ecosystem around their platform. But this is weak interoperability on a leash: the APIs are incomplete, throttled, heavily controlled, possibly unstable, and more importantly proprietary. To "let individuals download complete sets of their own records" falls into that category.


Strong interoperability is based on Open Standards (i.e. external modeling and specification definition by a broad consortium) and allows data portability (i.e. complete move to another service with reasonable ease, like switching broadband, mobile provider or bank). Open Standard creation is an iterative, imperfect, difficult and lengthy process because it is very detailed and requires negotiation from various parties with different views and interests. But is it any more costly or complicated than repeated enquiries by officials to address the excesses of monopolies? Is it more disruptive or complicated than breaking up large companies?

https://www.eff.org/wp/interoperability-and-privacy

I share EFF's view that strong interoperability requirements should be incorporated into market regulations. Because it aims directly at one root cause of some types of monopolies: the impossibility for anyone to compete once a service reaches a critical mass (e.g. even a giant like Google failed to threaten Facebook with its Google+ service) and consequently the lack of alternative for consumers (hence immensely reinforcing the power of the monopolist). It saps one of the main sources of power of the company. People can vote with their feet and mass exodus are possible. Something which is currently not possible with Facebook or Whatsapp without loss of contacts and fragmentation.

I am convinced that if the right mandatory interoperability mechanisms could automatically kick in early enough during the business growth this would be very effective at mitigating if not entirely avoiding monopolies way before the size and power of the company becomes a national or global issue. The advantages are that it happens before the critical mass is reached and it is a general mechanism as opposed to case-by-case anti-trust actions before the process starts all over again, and far too late, in 5-10 years.

One naive idea would be to cap the relative number of users a product or service can reach unless the vendor demonstrates adoption of Open Standards. It would greatly incentivise leading companies to initiate the dialogue process with regulators and other participants way before it reaches that threshold and thus absorb the costs but also deliver a public service by generalising, federating and formalising a diverse set of needs. The primary success criterion is that at any point it is possible for any company complying with the standards to serve a viable and sufficiently complete alternative to the dominant service/product. Users can jump ship, but still interoperate with the market leader, just like mobile networks, or mobile phones (thanks to 4G/5G).

Open standards can also be designed to be modular and enable granular third-party delegation. For instance recommender algorithms algorithms could be substituted or assembled from various parts.

The idea is to recognise that a dominant product/service becomes a public utility and as such has to be standardised, opened up, modular, interroperable, or entirely substituted without major loss or fragmentation for the users.

My point is that market diversity is probably the best and most natural antidote to concentration of power and actual empowerment of end-users. Governments have a important role to play but they are slow, dangerously influencable, and lack a solid comprehension of complex industries.
Anonymous said…
Just a rectification:

The AFP article that you link is from 2015. Amazon has, since, caught up:

https://www.cnbc.com/2019/10/24/amazon-announces-three-new-renewable-energy-projects.html

https://www.usatoday.com/story/money/2019/09/19/amazon-ceo-jeff-bezos-environmental-sustainability-emissions/2361987001/

https://www.businesswire.com/news/home/20201210005304/en/Amazon-Becomes-World%E2%80%99s-Largest-Corporate-Purchaser-of-Renewable-Energy-Advancing-its-Climate-Pledge-Commitment-to-be-Net-zero-Carbon-by-2040

https://www.zdnet.com/article/amazon-unveils-its-largest-single-renewable-energy-project-ever/