Skip to content
All articles

Why SustainAI Global Is Needed

Why SustainAI Global Is Needed (And Why Altruists May Be Missing Big Opportunities)

· Carter Sharp, Upskilling Associate · SustainAI.global

I want to preface this that I approach this with humility, both because of SustainAI Global principles and acknowledging that the ideas I present may be obsolete, incorrect, or otherwise not helpful despite my best intentions.

Three (flawed) stories about the future

Ask people who work on making AI go well what future they're aiming for, and you'll generally get one of three answers. For those who haven't read Senior Research Fellow at Forethought Research William MacAskill's "Viatopia", it boils the first two approaches down to:

The first is utopianism: figure out the ideal end-state ("end goal") for society, then aim straight for it. The main flaw is the hidden problems in their vision of the best future; there's simply too much they don't know to make a well-informed one. And, frankly, utopianism has a terrible track record.

The second is protopianism. Protopianism, coined by futurist Kevin Kelly, is "short term" by saying "let's not have an end goal at all, and instead take it one step at a time." The main flaws are (1) not having a clear idea of what to work towards means not having a better way of knowing which problems to be working on more than the others and when, plus (2) if we only measure success by what seems good in the short term, then that may come at the cost of long-term success (not having a clearer idea of the long term means we cannot have a clearer idea of if what we do now is better or worse for the long term—it's much less measurable). [2]

For clarity, I'm identifying a third existing approach: something in between that may involve thinking beyond the short term but not having a truly clear idea of what to work towards. Everyone who doesn't subscribe to utopianism or protopianism (or viatopianism) can generally be categorized here.

Viatopia: a better way

MacAskill's proposed alternative is viatopianism—focused not on the utopian end-state or the protopian next problem, but on the (as I think of it) medium term: a clear enough picture of the next position or waystation worth reaching. Once we get there, we set the next one, and the next, updating as we learn more about what the real destination should be.

Viatopia: "an intermediate state of society that is on track for a near-best future" — regardless of what that future ultimately turns out to be.

This gives us a goal to work towards for prioritizing problems, a way to measure success beyond the short term, and is less likely to be flawed than aiming more into the future. Most of all, it puts us in a better position to reach an actual utopia (or eutopia) because we'll update the next goal to work towards as we go as we learn more about what utopia to ultimately aim for—we're free to revise the destination as we go.

MacAskill doesn't offer a specific definition of viatopia—only an illustration, built around two categories of what I'm calling viatopian goods:

  • Societal primary goods: things essentially any future society would value, such as material abundance, scientific and technological capability, the ability to coordinate instead of fight, the ability to trade productively, and very low levels of catastrophic risk.
  • Steering goods (my term): the things that give a society the power to actually steer its own trajectory, rather than just survive it.

Where altruism is missing the full picture

As an example, we know from Jasmine Sun's reporting that those working at the biggest AI labs largely subscribe to utopianism. They believe that their work will lead to a utopia where (among other things) nobody has to work—but we at SustainAI Global argue that people should and need to be employed or face a crisis of personal meaning. One or both of us could be wrong, but either way, we can't know for sure until it's too late unless we're careful. Applying viatopian thought here could ensure we still have related steering goods that allow us to change course when we can make much more well-informed decisions. [4]

Many altruistic organizations are addressing pressing problems and doing wonderful work—but "very low levels of catastrophic risk" is only one of five of the societal primary goods MacAskill offers. Perhaps they are generally only focusing on what we want less of and what we should have more of, which is only a part of the equation for a good future. However, with viatopian thought, we can address the full range of things we should be working towards. Maybe that looks like those organizations staying the course but avoiding detracting from the other viatopian goods as they specialize in the "very low levels of catastrophic risk" good, while other organizations are created or pivot for the longer-term increase in all viatopian goods. What I think might be best is to have the general public hold larger organizations accountable for aligning their decisions with a collectively decided viatopian vision.

SustainAI Global: one application of viatopia

SustainAI Global is one implementation of MacAskill's viatopian ideas, not a new theory competing with them.

Among other things, it can serve as a one-stop shop for the masses wondering what to do while there's so much uncertainty around jobs in the age of AI. SustainAI Global's approach is to work backwards in order to provide a path forward in terms of what professional skills to develop:

Future society we want → organizations likely to be in it → jobs likely to be employed by them → their core skills.

That means we need a clear enough idea of the future society, so we created our own definition of viatopia. Following SustainAI Global's founder, Jacques Camier, in his work detailing Meta Quotient (MQ), I'm going to call our specific definition "MQS," short for Meta-Quotient Society, due to the alignment of Camier's explanation of MQ with viatopian ideas.

Why we're comfortable being different

SustainAI Global's projects won't always look like the projects of well-known altruistic organizations working on AI, but I argue that's a good thing.

Even if we apply viatopian thought in a way that is unique and not exactly what MacAskill and fellow researcher Jordan Arel outline in their public ideas around viatopia, that may help viatopian ideas spread. A couple related ideas are "parallel experimentation" and "natural selection of ideas/approaches," where we're collectively throwing spaghetti at a wall and seeing what sticks—which projects/approaches work, and which don't. While one could argue that we should just focus on different approaches in the same categories of approaches that viatopian experts recommend, I argue that there's also value in the idea of trying some ideas outside of that category, such as SustainAI Global's unique approach to identifying what skills to build proactively by working backwards from a vision beyond the short term.

This is important to note because MacAskill and Arel do not name general workforce education as one of their recommendations about how to go about practicing their viatopian ideas (MacAskill, How to Make the Future Better; Arel's Viatopia Interventions), but I still believe that it will be valuable. That can mean supporting their recommendations where SustainAI Global doesn't directly implement them, as well. [5][3]

Supporting the wider movement: the Good World Project

One great example of that is Jordan Arel's Good World Project. It's essentially having a chatbot walk the general public through coming up with their own vision of the "ideal future," and, by uploading each to a central database, they can be used to inspire more initiatives to improve the trajectory of that society.

SustainAI Global can easily integrate well with that idea in any number of ways beyond recommending people create their own vision—for example, taking their vision and recommending jobs that may or may not exist yet but that they can prepare for in order to work towards manifesting that vision, or using the collective input there to improve our MQS.

Another reason I point the Good World Project out is because it involves the general public more than Arel's other interventions—the keyword being accessibility.

Accessibility: a seat at the table for everyone else

If we take Arel's framing of "AI stakeholders" in this viatopian context:

  • AI labs (who builds AI)
  • Governments (who regulates AI)
  • The general public (neither, but who collectively have social and political influence)

...then we can categorize SustainAI Global as empowering the general public while organizations focused on making everything with AI go well tend to specialize in targeting the small number of people with leverage over the world's trajectory with AI, who are largely in AI labs and governments. [6]

Accessibility with making AI go well means allowing everyone to have a seat at the table—regardless of socioeconomic status and other factors that limit people's ability to have a say in AI affairs. SustainAI Global can create that accessibility through projects that give us, as the least utilized stakeholders, more say in AI affairs. It also means the other kind of accessibility with the products themselves—for example, by working around how time-poor people tend to be in the digital age.

This is important as another way of viewing SustainAI Global's more unique approach to making AI go well: focusing on the more neglected stakeholder group.

Routing talent and partnerships

If the general public uses SustainAI Global as a one-stop shop for everyone wanting to make more informed career choices, then SustainAI Global's ESI, Ethos, will be able to route talent towards resources and ideas personalized for their circumstances. That can mean recommending they visit other organizations who are focusing on AI labs or government and policy, but if that's not suitable, it will practice accessibility by meeting them wherever they're at and propose careers that require less formal credentials, and so on.

Nobody is left behind because everyone can come away with ideas of how to fit into, and create, a better future, specifically in terms of what skills to develop ahead of it and what jobs to aim for.

SustainAI Global can foster a great "coming together," where we use partnerships with other organizations to help unite our efforts toward the common goal of improving our world's trajectory. Much of that value may come from the "talent routing" mentioned where people are given workforce information and opportunities, but I believe most of it may simply come from giving people the chance to know about viatopia and have hope that we can collectively design the future we want for ourselves and our descendants.

Ending on a hopeful note

Even the people creating AI say that they don't know what career advice to give 17-year-olds, according to Jasmine Sun. The people who may have the greatest insight into the future capabilities of AI are effectively telling us there's good reason to be uncertain about the future of work. [4]

But what they don't seem to talk about is that we have the ability to create more hope and certainty by collectively making the future go in a better direction—such as by aligning our goals and trying to steer towards a viatopia.

A quote often misattributed to Abraham Lincoln is "The best way to predict the future is to create it." SustainAI Global seeks not just to predict what a better future looks like in the form of MQS, but to create it through action.

The invitation

If viatopian thinking resonates with how you think about AI, work, or the future more broadly, we'd welcome your involvement—get in touch.

Disclaimer: References to third-party researchers, projects, podcasts, or organizations (including William MacAskill, Forethought Research, Jordan Arel, Kevin Kelly, and 80,000 Hours) are provided solely for educational, analytical, and commentary purposes. Mention of these entities does not constitute or imply an official endorsement, affiliation, or sponsorship by or with SustainAI Global.

References & citations

  1. MacAskill, William. "Viatopia." Forethought Research, January 7, 2026. https://www.forethought.org/research/viatopia
  2. Kelly, Kevin. "Protopia." KK (Substack), 2026. https://kevinkelly.substack.com/p/protopia
  3. Arel, Jordan. "Building Cooperative Viatopia: Shortlist of Viatopia Interventions." jordanarel.com, October 21, 2025. https://jordanarel.com/2025/10/21/4-building-cooperative-viatopia_-shortlist-of-viatopia-interventions/
  4. Qureshi, Zershaaneh. "#248 – Jasmine Sun on what the people building AI really believe." 80,000 Hours, July 21, 2026. https://80000hours.org/podcast/episodes/jasmine-sun-ai-anthropology/
  5. MacAskill, William. "How to Make the Future Better." Forethought Research, 2025. https://static1.squarespace.com/static/5506078de4b02d88372eee4e/t/68ac54fdc7e83a183667629f/1756124413751/How+to+Make+the+Future+Better.pdf
  6. Arel, Jordan. "Viatopia and Buy-In." Effective Altruism Forum, 2025. https://forum.effectivealtruism.org/posts/CS7yTHe2qPbcXPLJb/viatopia-and-buy-in-1

Related SustainAI Global essays:

When citing this article: Sharp, Carter. "Why SustainAI Global Is Needed (And Why Altruists May Be Missing Big Opportunities)." SustainAI Global, August 17, 2026, https://sustainai.global/articles/posts/why-sustainai-global-is-needed/.