Kelvin Inobemhe
This year, when Senator Bernie Sanders made his appearance at the National Press Club to introduce the legislation to give the American people a 50 percent stake in the country’s largest artificial intelligence companies, he made his choice of framing very clear.
He did not speak much about rogue algorithms or existential risk, which have been a major part of the conversation on AI safety.
He discussed the concept of oligarchy.
He said that AI was created with the knowledge and creativity of millions of people, but the vast wealth was accruing to a small group of technology executives and investors.
His proposed American AI Sovereign Wealth Fund Act would impose 50% taxes on the biggest AI companies, and then have their stock held in a public fund, modeled somewhat after Norway’s oil wealth fund, while also granting an independent commission a vote on the boards of OpenAI, Anthropic and other such companies.
Sanders has been joined by earlier legislation he co-sponsored with Representative Alexandria Ocasio-Cortez, which focuses on the unchecked growth of energy-intensive data centres, and has a broader message he has taken on his “Fighting Oligarchy” speaking tour: Who owns AI?
This distinction matters. The debate that has been most in the public discourse in Washington, Silicon Valley and around the world has revolved around containment: will AI systems fool their users, can they be weaponised? Should they be held in check until safety research is completed?
It is a structurally different question, one with a longer history in American political economy, that Sanders is asking.
It was the question Louis Brandeis asked of the trusts a century ago, and each new iteration of concentrated capital has been asked – what happens to a democracy when a few firms dominate not just a market, but a whole technological system on which the rest of society increasingly relies?
The computer, data, capital and tools that displace labour, which are used by almost every sector of the economy, are now in the hands of a few companies.
Sanders says that is not a bad effect of innovation; it is what is happening right now in the world of AI, and it needs the same kind of scrutiny that trust-busters gave oil, steel, and rail.
The size of the scale gives the argument some form of credibility.
Today, a few companies, including some that have hit or surpassed trillion-dollar valuations, are in control of the foundation models that lay the basis for AI.
The cloud infrastructure that leverages them, and the chips that power most of the world’s AI applications.
Today, however, what gets built, who it is licensed to, and for how much, is being decided by a handful of boardrooms and, increasingly, by the operations of governments far removed from the borders of America, and their implications for labour markets and information ecosystems.
The proposal by Sanders to tap into that private money and use it to build a publicly held fund under the control of a commission with board representation is best seen not as an effort to publicise the money, but as an attempt to bring democratic accountability to this technology that until now has been largely governed by the market and by the firms themselves.
Sanders’ critique is grounded in the labour question, and more than simply hypothetical.
Here are some estimates of the impact of AI on jobs in the next 10 years: according to a recent report by the World Economic Forum, AI could displace as many as 40% of customer service roles and 70% of administrative positions over the next decade.
Meanwhile, a study by the Financial Times found that AI is poised to drive the replacement of as much as 10% of entry-level coding jobs in the next ten years.
Additionally, a report from the Brookings Institution projects that AI will replace between 20 and 30% of logistics roles over that same period.
And a recent McKinsey report estimates that AI could eliminate 80% of creative jobs, including those in the media, publishing, and entertainment sectors.
What worries Sanders and other lawmakers who are allies is the nature of the automation taking place this time, not so much its existence.
What is bothersome to Sanders and other lawmakers who have his back is not so much the existence of automation, but the conditions under which it is happening.
The earlier rounds of automation were negotiated, at least in part, either via unions or via public investment in retraining, or by decades of political wrangling about the distribution of productivity gains.
It is being used by companies with relatively weak organised labour opposition in the field of AI and which have, so far, benefitted from the increased productivity primarily in the form of profits for shareholders and not in widespread wages.
The idea of a sovereign wealth fund, of which Sanders’ is one, is a post-the-facto solution to that distribution.
What this means for developing countries is that concentrations of power are a problem for American workers, and for American democracy, but they are a more serious issue for the nations whose workers do not have a seat at the table on which these businesses are created, funded or governed.
Nigeria, like most of the developing world, is not a producer of frontier AI, it is a consumer and, to an extent, a subject of decisions made in Washington, Beijing, San Francisco and Brussels.
From the computing power to the foundation models, the cloud contracts, to even the data sets to train systems that Nigerians increasingly depend on for banking, agriculture, healthcare triage and even government service delivery, nothing is in their own hands.
This is the developing-world version of Sanders’ oligarchy argument: Where a few companies are able to exercise a lot of power in the United States, the same set has far more power when multiplied by the fact that there is little regulatory power, less capital to compete, and less capacity to influence the rules of the game in the developing world.
Nigeria’s stake in the argument is quite significant in the sense that to Nigeria, it is apparent in several specific ways.
The same pressures on jobs that Sanders talks about for American workers hit Nigeria’s large youthful, and heavily informal, labour force especially hard, with few safety nets, retraining facilities and social insurance to buffer against the onslaught of automation. In digital inequality, the divide is not just between those who have broadband access and those who do not, although that is a major factor; it is also between those who possess the compute and capital to create AI systems, and those who do not, who are, for now, almost entirely dependent on foreign cloud providers and foreign-developed models, and whose visions can only be enabled through them, and whose terms are set by them.
Regarding data sovereignty, Nigeria has recently enacted the Data Protection Act of 2023 and the Nigeria Data Protection Commission, which is still in the process of being established, are genuine and comparatively early steps in asserting some degree of domestic control over the collection and use of personal data; and the NDPC’s stated intention to look at the Act specifically with respect to AI reflects recognition that the rules were not drafted with generative AI in mind.
However, data protection law only covers one of the sovereignty issues; it does not alone solve the bigger question of the ownership of the infrastructure and models that are created on top of the data, which are most on servers and under corporate ownership frameworks, well outside Nigeria’s jurisdiction.
Concerning regulation, even though Nigeria does not have a dedicated AI Act, there are sector-specific guidelines provided by the Nigerian Data Protection Commission (NDPC), the Nigerian Information Technology Development Agency (NITDA), the Central Bank, and the Nigerian Communications Commission (NCC), in addition to a National AI Strategy in its final stages.
Responsible innovation and capacity building locally is a good starting point, but not yet a binding, enforceable contract that will give Nigeria leverage over the use of foreign AI firms in Nigeria, pricing of their services, and training of systems that they do not derive concrete benefits from, by using Nigerian data and Nigerian languages.
None of this should be interpreted as a call for Nigeria to emulate Sanders’ sovereign wealth fund model literally; it does not have the kind of tech giants that are involved, and does not have to be the same kind of tax base to allow it to such an extent.
The principle that the governance of a transformative technology should not be solely in the hands of firms that benefit from it, however, directly converts into a Nigerian policy agenda.
The agenda should feature a genuine, enforceable framework for AI governance with regulatory bite, not just guidance; deliberate investment in the local compute and research capacity of Nigeria, rather than continuing to be a data source and a market; data localisation and data sovereignty requirements that take into account Nigeria’s informal sector and large youth population; and labour policies that account for the specific dynamics of Nigeria’s informal labour market and youth population, rather than being copied from wealthier countries.
At first glance, Sanders’ battle is a home-grown domestic American struggle against oligarchy and inequality.
But the question that has never been more obvious- whether a small group of businesses are sufficient to define the future of a technology that could have such widespread implications, without input from the people- is not uniquely American.
But for Nigeria, the consequences of being wrong, if any, are even more dire, given the little influence that the country has in the writing of the answer elsewhere.
Inobemhe is a communication and media researcher. He writes on media, democracy, and governance in Nigeria
For a better society
Follow us across our platforms:
Instagram – https://www.instagram.com/championnewsonline/
Facebook – https://web.facebook.com/championnewsonline
LinkedIn – https://www.linkedin.com/company/champion-newspapers-limited/
https://x.com/championnewsng/
You can also like and comment on our YouTube videos.
https://youtu.be/QIBfD1tT80w?si=R4Qf3so2LxYu3GC2
