「笑える」世界 AI 専門家がザッカーバーグの「AI for everyone」に異議

On August 10, Mark Zuckerberg published a letter titled “The Future for Everyone,” a manifesto laying out his vision for how superintelligence could improve people’s lives, expand human capability, and distribute the benefits of artificial intelligence more broadly.

Zuckerberg advocated for wider distribution of AI access across countries and companies as “the path to a positive AI future,” sharing ways in which Meta is trying to make it happen.

0:00/8:22

*Rest of World* spoke to six AI observers from around the world about what they made of the letter, and whether AI can really become the great equalizer Zuckerberg envisions.

Some asked who gets to decide what “everyone” means. Others highlighted how access doesn’t always equal power.

Their responses have been edited for length and clarity.

“Grossly overstated economic benefits”

Chinasa T. Okolo, founder and scientific director of policy incubator Technecultura, and the inaugural policy lead of research organization Black in AI

AI infrastructure investments in African countries have yet to yield any significant tangible benefits. Kenya is pausing its Microsoft and G42-backed geothermal data center due to grid constraints, and forthcoming efforts like Cassava AI’s Nvidia-powered AI factories are likely to exacerbate grid stability concerns in countries like South Africa and Nigeria.

Meta has grossly overstated the economic benefits of its data centers.”Chinasa T. Okolo, founder and scientific director of policy incubator Technecultura

Meta has grossly overstated the economic benefits of its data centers, not only in this manifesto but also through extensive advertising campaigns focusing on smaller rural towns that traditionally hosted manufacturing or mining plants that have largely gone bust.

The labor benefits of data centers are primarily concentrated during construction, and once operational, they employ only a small fraction of permanent workers. While there is a marginally increased impact on home prices, this benefit doesn’t translate to local wages. Data centers are also associated with negative health and environmental impacts, which are likely to be even more challenging in African countries, which have fewer protections and lack the ability for governments to oversee strict compliance if environmental laws exist.

“Excludes people at the margins”

Sagar Vishnoi, director and co-founder of Future Shift Labs

Zuckerberg’s definition of “everyone” largely excludes people at the margins of the digital ecosystem: speakers of low-resource and tribal languages, shared-device households, communities with intermittent connectivity, and populations for whom digital infrastructure remains uneven.

If companies genuinely want AI to educate billions, investment cannot stop at bigger models. It has to extend to building open language infrastructure — consent-driven data sets, community-led language digitization, and compute access for local researchers and developers.

Trust will ultimately depend less on how much data AI can collect and more on whether platforms can demonstrate accountable enforcement, independent oversight, and protections that work across diverse legal and social contexts — not just within Silicon Valley’s institutional framework.

“Overlooking a fundamental reality”

Mark Mulobi, products and practices associate, Digital Impact

As a Kenyan, what resonates most with me in this vision is the potential for AI to act as a personal tutor — helping people learn, build skills, and access new opportunities. But the promise of “AI for everyone” risks overlooking a fundamental reality: Not everyone is online, digitally proficient, or able to engage with AI in English.

Realizing AI’s potential in Kenya — and across Africa — requires confronting the last-mile realities that still shape who can participate: reliable electricity, high-speed internet, affordable devices, and meaningful access in rural and underserved communities. It also requires deliberate investment in African languages and local knowledge systems. Without those investments, AI risks widening existing digital divides rather than becoming the empowering, inclusive technology that this vision aspires for it to be.

“A vector of U.S. power”

Juan Ortiz-Freuler, co-founder of global civil society coalition Non-Aligned Technologies Movement

Meta is so big it runs afoul of many different interest groups. These contradictions become more acute as the U.S. expects its private multinational tech corporations to play a key role in upholding U.S. hegemony, which often contradicts its profit-driven motives.

Meta is so big it runs afoul of many different interest groups.”Juan Ortiz-Freuler, co-founder of global civil society coalition Non-Aligned Technologies Movement

So while the first sections of the document target the global community — and regulators in places like Brazil, India, Turkey, and Nigeria — with a message of transparency, horizontality, and willingness to cooperate broadly, the latter sections are directed to U.S. regulators and power brokers who expect Meta to be a vector of U.S. power.

While Facebook is not a data-center company, it is an intelligence company. Its profit interest here comes not with increasing compute but with relying on humans to help untangle the relations between people and things. … Essentially, Facebook is redistributing tasks to its users across the world while training a central machine to understand the complex web of relations that underlie the world of symbols it was fed as text. It will use this to optimize its machines for a variety of tasks (curation, targeting), which it acknowledges it will also put at the service of the US government and the America First agenda.

Abeba Birhane, founder of Trinity College Dublin’s AI Accountability Lab

Zuckerberg’s claim of “AI for everyone” is nothing new.

Decades ago, when Facebook overtook much of the African continent, Zuckerberg’s motto was something like, “connecting the rural and unconnected women across the continent.” Now, with a full grip on the continent’s digital infrastructure, Facebook/Meta has amassed populations’ data, monopolized the digital sphere, and controls the information ecosystem, with no observable benefit to the actual rural women it once boasted it would benefit.

Facebook has since exploited cheap or unpaid labour of data workers, extracted our resources, and caused enormous harm through disinformation, misinformation, and malinformation, while the corporation’s response when it is called out has been to displace accountability at any cost and gaslight the very people it claimed to help.

Looking beyond Africa, nobody should take Meta or Zuckerberg’s letter to “benefit everyone” seriously when the corporation that continues to silence criticism, retaliates against its own workers for exposing its technology’s harms, buries evidence of harm to minors, actively launches problematic products (the perv glasses) when overworked and underfunded civil society organizations have their attention focused elsewhere, and continually violates its legal obligations somehow claims to have people’s — especially vulnerable groups’ — best interests at heart. It’s laughable.

“Many aspects of trust that this letter omits”

Priya Vora, CEO of global nonprofit Digital Impact Alliance

Technology must be grounded in four guiding principles: participation, agency, choice, and trust. The letter speaks to participation, but meaningful participation requires people to have agency and choice — and to trust the systems they are being asked to use.

Governments can establish the rules of the road, but companies have to earn people’s trust through their actions. That means designing AI hand -in -hand with users — especially with those for whom accessibility needs are high — being transparent about how data and AI systems are used, and, crucially, giving people meaningful recourse when AI gets something wrong.

There are many aspects of trust that this letter omits. Trust is multidimensional, and those dimensions need to be considered from the outset and throughout the development and deployment of AI if we want these technologies to truly work for

everyone.