August 20, 2026 9:01AM # Manifesto Destiny: Can Meta Deliver on Individual Empowerment? By [Kevin T. Frazier](https://www.cato.org/people/kevin-frazier) --- SHARE Listen to this article Generated with [ ElevenLabs](https://elevenlabs.io/) AI technology. Loading the [Elevenlabs Text to Speech](https://elevenlabs.io/text-to-speech) AudioNative Player... ![Artificial Intelligence](/sites/cato.org/files/styles/pubs_2x/public/2024-06/artificial-intelligence-1600x900.jpg?itok=DgCb5p97) (Getty Images) In the so-called [AI manifesto war](https://www.axios.com/2026/08/02/ai-manifesto-open-weight-models), Meta CEO Mark Zuckerberg has adopted an offensive posture that stands out from those issued by his peers. “We propose a [philosophy based on individual empowerment](https://about.fb.com/news/2026/08/the-future-is-for-everyone/) as the source of prosperity, invention as the primary purpose of superintelligence, and balance of power as the foundation of safety,” he writes. Whether Meta will practice what it preaches—and spur others to do so—is an open question. **A Call for Individual Empowerment** Zuckerberg translates this philosophy of individual empowerment into policy recommendations that facilitate ongoing AI development and widespread AI adoption. Rather than allowing the federal government to screen frontier AI models for [up to a month](https://thehill.com/policy/technology/6017847-trump-closed-door-ai-framework-withheld/), Zuckerberg proposes a more flexible process that would grant regulators access to models at earlier stages of development—enabling the government to take any necessary steps to safeguard systems without stalling the ability of labs to release models expeditiously. He likewise emphasizes the importance of supporting open-weight models. Ready access to open-weight models will, per Zuckerberg, allow individuals to create AI tools tailored to their interests, values, and needs; empower researchers to improve the science of AI; and counter trends toward the concentration of power. Aligning with other manifesto authors, such as then–[Google DeepMind](https://x.com/demishassabis/status/2076957440109625718) [CEO Demis Hassabis](https://x.com/demishassabis/status/2076957440109625718) and [AnthropicCEO Dario Amodei](https://www.anthropic.com/news/position-open-weights-models), Zuckerberg acknowledges the China-shaped elephant in the room. He urges local, state, and federal stakeholders to assist with an ongoing AI infrastructure build-out as China and other, unnamed countries allegedly add over one gigawatt of nuclear capacity every other week. Other labs should learn from Meta’s policies and history, in Zuckerberg’s estimation. His review of history reveals several foundational policy principles. “There is no single objective answer to how people define the best life.” He regards this as the reason to make AI tools as broadly available as possible while also empowering users to determine how, when, and why they use such tools. “\[H\]oping that an absolute power will benevolently provide for humanity if sufficiently enlightened has not led to safe or positive outcomes,” he states, calling to resist policies that entrench today’s AI leaders. He further states that “there is no rule that AI must increase automation faster than it increases individuals’ capabilities or demand for new skills,” which cuts against forecasts that AI will lead to severe economic disruption. He also thinks that other labs should empower an independent board to “approve the safety criteria for releasing models and review\[\] whether each model release adheres to the criteria,” as Meta does. The overriding message of the manifesto is that AI can and should be a tool of individual empowerment. He foresees a world in which people leverage AI to be entrepreneurs, to be more present with family and friends, and to explore their personal curiosities as deeply as they’d like. Yet, it’s too early to tell if Meta has the will and, perhaps more importantly, the resources to make good on this ambitious, positive vision of the future. **Is** ***This*** **the Time Meta Follows Through?** Meta, then Facebook, has claimed the motto “For users, by users” on prior occasions. It launched a [mechanism for Facebook users](https://about.fb.com/news/2009/02/facebook-opens-governance-of-service-and-policy-process-to-users/) to vote on some platform policies. At the time, Simon Davies, then director of Privacy International, celebrated, “This is an unprecedented action. No other company has made such a bold move towards transparency and democratization.” The bold move [did not result](https://www.politico.com/story/2012/11/facebook-proposes-to-end-voting-on-privacy-issues-084167) in enduring change; the program was brought to a close unceremoniously. A few years later, Zuckerberg [concluded that](https://about.fb.com/news/2020/05/welcoming-the-oversight-board/) “Facebook should not make so many important decisions about free expression and safety on our own.” He then launched the Oversight Board, an independent body that reviews specific content moderation decisions and offers policy reforms to Meta. It has [largely fallen short](https://www.platformer.news/meta-oversight-board-5-years/) of expectations. And, looking to this manifesto, it’s difficult to reconcile Zuckerberg’s vision for a personalized superintelligence with “strong privacy and security options so you can trust it to handle all of your personal content knowing that no one else can access your information” while the [company draws headlines](https://www.wired.com/story/meta-accidentally-let-employees-access-each-others-keystroke-data/) for failing to adequately protect employee data. For Meta to live up to the manifesto, it will have to commit resources, alter policies, and reform its culture such that the failures of the past do not undermine this positive vision for the future. Of course, realizing Zuckerberg’s broad ambitions in many ways relies on the entire AI ecosystem—including labs, local, state, and federal officials, and the general public. The odds of those other actors rallying behind Zuckerberg’s aspirations, however, hinge at least partially on the lab leading by example. That said, regardless of whether Meta pursues this path, it’s pivotal that the AI ecosystem include actors that place a heavy emphasis on building useful tools for diverse audiences, with an eye toward advancing human flourishing. This mission is distinct from that of labs that have opted to predominantly build tools for large enterprises. For AI to live up to its hype, there must be a range of ways we allow innovators to develop tools beyond those simply high in demand by those with the most resources. This will give companies and individuals many more AI tools and allow individuals to adapt them to serve their individual well-being. **First, there’s the matter of AI access.** Zuckerberg is counting on a wide range of people—average Joes and Janes—making use of highly capable AI systems. There’s a long way to go toward that reality: [Fewer than 3 percent](https://www.pnc.com/content/dam/pnc-com/pdf/aboutpnc/EconomicReports/consumer_health_check/PNC_Research_Consumer_Health_Check_April_2026.pdf) of American households have paid subscriptions to AI tools. Many more make [use of free AI tools](https://www.pewresearch.org/short-reads/2026/03/12/key-findings-about-how-americans-view-artificial-intelligence/), but not for the sort of tasks, [such as deploying agents to perform extensive workflows](https://www.gallup.com/workplace/712736/organizational-adoption-jumps-six-points.aspx), that Zuckerberg expects. Unless and until the public knows how to put AI tools to full use and has the resources to do so, Zuckerberg’s vision will remain the stuff of blog posts rather than history textbooks. Meta plans to continue producing open-weight models, which substantially lower barriers to entry and may drive down the cost of access to closed models, but that’s far short of a meaningful fraction of Americans taking advantage of the superintelligence Meta expects. **Second, there’s the question of agency.** Zuckerberg is excited by the prospect of users designing their own AI tools, specifying the values they wish their models to follow, and regularly engaging AI in ways that further individual well-being. This echoes the vision set forth for the social network era. [A different letter by Zuckerberg](https://www.theguardian.com/technology/2012/feb/01/facebook-letter-mark-zuckerberg-text)—this one penned in 2012—asserted that “we \[Facebook\] build tools to help people connect with the people they want and share what they want, and by doing this\[,\] we are extending people’s capacity to build and maintain relationships.” He specified that “personal relationships are the fundamental unit of our society.” Then, as now, Zuckerberg framed technology as a means for people to identify and exercise agency. Instead, the promised social network transformed into social media. This transformation coincided with (and according to some, caused) a significant loss of agency. Individuals [spend more time online](https://www.bls.gov/opub/ted/2026/americans-age-15-and-over-read-for-an-average-of-16-minutes-per-day-in-2025.htm) than ever before, and [Americans are socializing less](https://www.axios.com/2026/07/05/americans-socializing-decline), marking a loss in Zuckerberg’s own metric of the fundamental unit of society. It may be the conscious and deliberate decision of Americans to allocate their time to online activities, but that time does not seem to contribute to [human flourishing or individual well-being](https://www.theatlantic.com/magazine/archive/2025/02/american-loneliness-personality-politics/681091/). Something has to give. While some people may fully tap into the potential of the internet and related digital tools to launch businesses, participate in civic affairs, and learn new topics, time online is generally time spent [scrolling social media feeds](https://www.broadbandsearch.net/blog/average-daily-time-on-social-media). It is unclear how to better align the interests and goals of users with their interactions with AI. **Finally, and relatedly, there’s the question of the direction of AI use.** Zuckerberg makes an economic case that because invention is more valuable than merely automating work, more AI use will go toward the former than the latter. This assumption is built into how Meta plans to set up user access to its tools: For those who want to pay to use more compute, there will be a dynamic auction mechanism that will guarantee that everyone gets the lowest price possible for the intelligence and compute they’re using while also ensuring the capacity is used for whatever people collectively find most valuable. This will ensure the benefits of superintelligence are distributed widely. The market does not seem to agree. Venture funds [directed toward AI agents](https://www.wsj.com/tech/ai/saas-software-as-a-service-apocalypse-ai-b9b6da99) for various professions, from assistants to zoologists, are off the charts; they are chasing [$1.4 trillion per year](https://ctl.mit.edu/news/mit-center-transportation-and-logistics-launches-ai-labor-exposure-map-quantifying-14-trillion) in wage-bill-equivalent work that AI may be able to automate. Meta has more work to do to orient AI investments around inventions to achieve Zuckerberg’s hope. **The Challenge and Opportunity Ahead for Meta and the AI Ecosystem** Meta and other developers of open-weight models are poised to challenge the undue influence that Anthropic, ChatGPT, and Google wield over public policy and the AI space generally. Zuckerberg has positioned the company to lead the open-source efforts here in the United States. He has rightfully contested the idea that a handful of unrepresentative individuals at an even smaller set of AI labs should dictate how this technology spreads. His manifesto outlines several thoughtful yet abstract goals that align with liberal values and carry the potential to unleash human flourishing. The question is what he is going to do to ensure Meta does not miss this opportunity to offer an alternative, truly open approach to the entire AI development and diffusion process. Advocates for this perspective—an open AI ecosystem that centers individual and collective human flourishing—should not put their eggs squarely in Meta’s basket, but Meta’s manifesto reinvigorates an important discussion on why open products should be part of the conversation. The AI ecosystem has many different approaches, not only from current leaders but from smaller players as well. While these differing visions are important, what is perhaps more important is that the government not choose a single vision or approach as the winner. Instead, different products and visions may teach us different things about the technology’s potential and serve different needs. Manifestos are easy to write and much harder to act on. Meta’s own history proves that bold declarations of user sovereignty can quietly expire. The measure of this one will be whether Meta spends, builds, and reforms like it means it and how such an approach allows not only Meta but others to compete in the market. ##### Related Tags [Technology and Privacy](https://www.cato.org/technology-privacy) [![Creative Commons License](/build/cato_2020/images/creative-commons.svg)](http://creativecommons.org/licenses/by-nc-sa/4.0/) This work is licensed under a [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License](https://creativecommons.org/licenses/by-nc-sa/4.0/).