Ignore the Big Tech Propheteers: Counter Thoughts on an AI Future

As big tech bosses weigh in on how artificial intelligence (AI) will change the world, I urge us not to take any particular path for granted and focus on the deep contradiction within capitalism that is emerging.

We have a new literary genre: Big Tech’s leaders, former leaders and uber-investors telling the world how the future will and should be. Dario Amodei’s recent call to ‘pace’ frontier AI development is just the latest example. Call it Big Tech ‘propheteering’.

The era of GenAI has seen an explosion of this genre, with venture capitalist Mark Andreesen’s 2023 techno-optimist manifesto; ‘Gentle Singularity,’ the essay by OpenAI’s Sam Altman; the 2024 and 2026 essays by Anthropic’s Dario Amodei on the promise and risks of AI, respectively; and the notorious X-published manifesto by Palantir’s Alex Karp from March this year. The light-news days of August 2026 saw two new additions: Meta founder Mark Zuckerberg’s paean to ‘superintelligence’ and Microsoft founder Bill Gates’ more sombre essay on ‘the turbulent AI era’. 

Zuckerberg is no stranger to the propheteering genre, having in 2017 proclaimed Facebook as the global community connector, just when its brand reputation needed polishing. But his latest essay, ‘The Future Is for Everyone’, is notable for many reasons.

‘Open source’ AI?

First, because the focus has shifted from Meta’s earlier business models (social media, then the hardly-lamented Metaverse) to its latest: its own ‘open source’ AI. This is the ‘superintelligence’ whose transformative effects Zuckerberg proclaims. While the essay at first sight offers a general vision of AI’s gift to humanity, it is a gift that comes with strict conditions.

According to Zuckerberg, allowing closed corporate monopolies to exploit AI frontier- models will prevent superintelligence’s benefits being distributed as widely as humanity needs. Indeed, he argues in a neat rhetorical twist, companies such as OpenAI and Anthropic are in bind: insisting both that their AI is of huge benefit to all humanity, but also requiring they retain control over its release and development, because of its innate risks. There is indeed a contradiction here, but what’s entirely unclear is whether Meta’s ‘open source’ AI is the solution.

Even if we leave aside the issue of in what sense Meta’s AI really is open source, Zuckerberg’s celebration of superintelligence’s benefits—a super-powerful personal agent and PhD-equivalent personalised tutor for everyone, universal access to science, entrepreneurial opportunity and ‘invention superpowers’—is strikingly silent on the business side of Meta’s new offer. There is no mention of Meta’s parallel attempts to exploit its computing power. Even more significantly, there’s no mention of Meta’s plans to exploit its AI users as a revenue source, whether as a means to drive consumption from third parties or as eyeballs for its WhatsApp business customers.

The overall goal of Zuckerberg’s prophesying is clear: to legitimate, or at least sanitise, the pivot to a new business model (without actually mentioning it) that Meta’s increasingly toxic public standing requires.

‘Democratising’ AI

This takes us to the second interesting point about Zuckerberg’s essay: the contradictory nature of its overall vision. In PR terms, as just mentioned, the general goal is clear: to reposition Meta as the friendly, ‘democratic’ version of Big Tech’s AI offer, just as public suspicion of that general offer is rising steeply.

The ‘democratizing’ move is not new: Satya Nadella, Microsoft’s CEO, has talked of its AI ‘empowering’ people for a long time. But Zuckerberg makes it fully explicit. Meta’s AI tools will ‘put[] power in people’s hands’; indeed ‘the future will be created by all of us’, because everyone will have access to the same extraordinary AI tools. I’ll come shortly to whether that idea is even coherent, but let’s stay on the surface of the argument first.

The very idea of ‘superintelligence’ ignores completely the questions raised by philosophers like Shannon Vallor about whether AI’s intelligence really is the same as humans’ embodied intelligence. To call AI a ‘superintelligence,’ one that only needs to be made available to humans in order to benefit humanity, implies that it is just a lot more of the same ‘stuff’ that humans have. But it isn’t: as the constant scandals about AI scheming bring out, machine ‘intelligence’ is very differently grounded from human intelligence, if it is grounded at all. All the potential social costs of deeply embedding machine intelligence into the world are conveniently ignored by the simple notion of ‘superintelligence’.

Meanwhile, in a strange paradox, Zuckerberg’s vision that ‘the future is for everyone’ is combined with a proposal to use Meta’s vast cash-piles to subsidize local communities, while ensuring its future computing infrastructure actually gets built. Zuckerberg proposes a ‘The Future is For Everyone’ (sic) fund to subsidize AI training and bonus payments in communities where Meta’s exclusively controlled data centres are supported, managed through ‘strong community compacts’. As if the answer to the growing AI rebellion was pseudo-feudalism: local pacts with a friendly infrastructural overlord.

Market liberalism

Zuckerberg’s lapse into feudalism is, however, in tension with the essay’s underlying driver: a zealous belief in market liberalism, the third notable point about the essay. The main explanation he offers in favour making AI ‘for everyone’ is not a vision of democracy at all, but a vision of perfectly functioning markets.

At various points, Zuckerberg extols the ‘natural balance’ between companies ‘working to serve people’s needs’ and ‘individuals’ becoming more empowered. The idea, in the essay’s words, is of a ‘natural economy of ‘natural checks and balances’ which—somehow naturally!—will overcome the short-term risks of unemployment from automation. Once again, AI’s social costs magically disappear in the language of market individualism: ‘the more superintelligence serves as a tool of invention, the more likely that individual capability outpaces automation and the future is better for people’. 

At the end of the essay, Zuckerberg theorises the balancing forces of individual empowerment as ‘recursive self-improvement’ (echoing a technical term from AI discourse). He acknowledges the tensions that computing powers as large as ‘superintelligence’ involve, but suggests that all will be fine ‘as long as we maintain a balance of power that favours power overall’. For, after all, he writes, ‘competing interests are natural’.

Liberal market individualism has rarely been as naked or naïve as this: any idea that competition is harsh and leads to power concentrations (indeed of the sort the essay said it opposes) somehow is absent when ‘the people’ are empowered. Will there then be no winners and losers in the age of democratized entrepreneurship?

We only had to wait two weeks for a partial rebuttal of this magical thinking, delivered  by a US West Coast boss of a different generation—Bill Gates, the Microsoft founder. In his own essay, Gates nodded to Zuckerberg’s vision (without directly referring to it): ‘AI will either be the greatest equalizer ever invented, or the worse source of injustice’. But his emphasis throughout was on the latter. For Gates, the problem is that AI, because it ‘can replace and even exceed human cognition’, will, through the operation of the very market forces Zuckerberg extolled, eradicate huge swathes human jobs without hope of replacement, unless there is a major social corrective.

Gates’ problem is very different from Zuckerberg’s. He does not need to massage the public reputation of a tech corporation—as he emphasized, he is now in the business of giving away his original billions—and so can focus on humanity’s actual needs. What he suggests humanity needs is a plan: a plan to address the huge social costs ignored in Zuckerberg’s essay, as currently (in Gates’ words) ‘there is no plan to ease the entry into the AI era’.

Is the ‘AI era’ inevitable?

While Gates’ proposal to maintain a ‘reserve’ of human-only jobs has rightly attracted attention, his initial formulation signals the enduring problem. Once we talk about ‘eas[ing] the entry into the AI era’, we have already assumed that era is inevitable, indeed in large part has already happened, that the ‘AI era’ is now. 

Yes, Gates does acknowledge a role for ‘a public democratic process’ about managing the AI risks, but should this be on the basis that widespread AI adoption is already inevitable? What neither Gates nor Zuckerberg, nor any other Big Tech boss past or present, can contemplate is that the peoples of many countries may have different ideas.

The AI future (and present) is only inevitable and stable if citizens and workers fail to see the serious choices about AI adoption that are, in fact, before us. This is surely the time to ignore Big Tech propheteering and begin seriously debating the real democratic choices over AI’s future.

Nick Couldry is the author of Predatory AI: How We Can Resist Corporate Capture of the Human Mind (Polity November 2026). Blog originally published at https://blogs.lse.ac.uk/medialse/2026/09/17/ignore-the-big-tech-propheteers-counter-thoughts-on-ais-future/. This post gives the views of the author and not the position of the Media@LSE blog, nor of the London School of Economics and Political Science.

Featured image: Photo: Wikimedia Commons