 |
| Don't cast out the giants; just help them better help us all. That's what governance and regulation are really about. |
But evidently the USA's market capitalisation is going through the roof. However, that's only due to about 7 companies – aside from them, the US stock market has actually been in decline for years (last I checked). Given what inequality (AKA power concentration) does to democracy, I was very happy to hear Macron say last November's Digital Sovereignty meeting in Berlin, that Europe doesn't want the "Seven Manifiques" taking over our democracies. But he still seems obsessed with "hyperscalers."
What are hyperscalers really? They are
- natural monopolies – that is, adding more customers or selling more products costs a flat amount. James Bessen has shown this is basically all companies of scale now, because they can use AI (or he just says "software") to solve logistics problems,
- able to access customers globally (not bounded much by geography or governments),
- which people have come to rely on broadly for their lives and work.
Or in other words (specifically, those of US antitrust law), hyperscalers are utilities.
People think antitrust law is about breaking up big companies, but that's not true. It's about enforcing higher standards on more powerful companies, and ensuring that new companies can enter markets. When there are good reasons that there will be few companies in a market, then the government has to at least partially replace the role of competition in that market in helping ensure quality, fair costs, fair wages, and so on. There is therefore an obligation on governments to:
- identify new utilities,
- strip any business components of the companies providing those utilities that could be regulated by the market off the utilities and spin them off into ordinary companies to compete in the regular market,
- set up a dedicated regulatory body for ensuring quality, fair costs, fair wages, and so on.
But the US hasn't done this since about 1971 – which to me indicates antitrust law was seen at least in part as being about redistribution and dealing with Marx' dangerous ideas, since that's about the time when the Soviet / USSR's economy plateaued. And in fact this could well be why Boeing is now having trouble with its quality – in 2001, George W. Bush didn't see why one company (the last commercial airline builder in the US) needed it's own independent regulator, so shut that down and used the resources saved to create the Department of Homeland Security. It's impressive really that Boeing kept its quality together almost two decades, because external regulation really simplifies governance. (I'm remembering a senior Googler's face when I explained this recently in a small room…)
So that's what needs to be done with big tech and various other sectors (pharma? consultancies? insurance companies?) showing consolidation. But the US isn't doing it, and I've heard recently can't do anything at least at the national congressional level at least because of the amount of money flowing to legislators. But they'd like to. They like everyone would like to see these companies contributing more for the populations and infrastructure that supports them.
But does it matter if the US wont' do this? Are these companies really American, or is the US just where they're domiciled? I realised that I think not, during COVID, due to Eric Schmidt. I was participating in a Zoom (or similar) where he was talking to German regulators in early 2021, and heard him say "Too bad you [Germany – the GAFAM tend to be anti-realist about the EU in an effort to make the Brussels Effect go away, even though they've actually found out the GDPR and most likely also the DSA and other legislation have helped them more than they cost in terms of compliance] Too bad you don’t have your own AI to regulate."
I immediately jumped into the chat to say this was wrong, for two reasons. Secondly, each state signatory to the Universal Declaration of Human Rights (UDHR, that is, all states) have an obligation to proactively defend the human rights of every human within their borders. But firstly, I don't believe we are talking about US AI.
- This AI is produced by talented people from all over the world (who have chosen to migrate there);
- often using the data of people all over the world (who mostly have not migrated, nor consented to their data use.)
- The data is gathered and the AI products and services transmitted using infrastructure local to those people.
In my opinion (which I've been talking about for years, including in UN fora), we need transnational utilities to be regulated transnationally. I would like to see a treaty organisation set up to do the work and handle the redistribution. We need to innovate and to iterate the kind of trade treaty – the EU os pretty good, but we need a better strategy than what we've seen in the EU or the UNSC for this kind of thing, e.g. to allow illiberal governments to drop out of compliance (and redistribution) rather than allow them to veto. One suggestion my colleague Malikova is now putting more of her time into is disaggregating these companies by regulatory region – I've also been talking a bit about that for years, but I think this might be one thing that really global (even UN) coordination would make sense for.
This is a core part of my book, but I took some time out from a great conference, DIGHUM, after I said the above there again, and some interesting people were hovering around me afterwards, which I realised in retrospect might have been there partly because I said some similar things in small rooms at GLOBSEC. So it's better to have all this material in one place where hopefully it is clear and undestorted. But let me know if anything is unclear, preferably here in the comments.
Comments