Many thanks to Max Taylor, Alistair Stewart, Albert Didriksen, Jeff and Johannes Pichler for feedback on this post. All mistakes are my own. This post does not necessarily reflect the views of my employer.
Executive Summary
- I believe that AI development could have an outsized impact on animal welfare and that these stakes warrant deep investigation. However, the complexity and uncertainty involved make it difficult to approach this.
- This post outlines three robust anchor points — concepts that we know and that I've found helpful — for navigating the uncertain and high-stakes intersection of AI development and animal advocacy. By focusing on what we do know, I hope to nudge and provide structure to more strategic research. These anchor points are:
- Influence and Values: Fundamentally, we want to increase pro-animal values in influential AI-related spheres and amplify the influence of the pro-animal movement itself, through both upskilling animal advocates on AI and strategic engagement with key AI stakeholders.
- Interest and Alternatives: Fundamentally, we want to increase interest in alternatives to factory farming and ensure viable, compelling alternatives exist, by leveraging AI to improve alternatives, increase interest in them, and dismantle systemic obstacles to their adoption.
- Symmetries and Asymmetries: Fundamentally, we want to recognize that the pro-animal movement is not alone in leveraging AI (as the animal agriculture industry will also deploy these tools), suggesting a strategic focus on identifying and exploiting comparative advantages to secure distinct benefits.
- Each anchor point guides toward key strategic questions, with examples presented for the first two points. The third point is discussed in a separate post.
Introduction
My key motivation to work at the intersection of AI and animal advocacy can feel abstract and uncertain. Ultimately, I believe something along the lines of:
- Over the coming decade(s), AI is likely to have an enormous, perhaps unparalleled, effect on our society, which will likely include implications for animal welfare.
- This introduces risks and opportunities that may be at a scale larger than any previous development in our world. We may have an opportunity to end factory farming within a few decades, but we are also facing the risk of the industry doubling (or growing even larger) in scale.
- These immense stakes warrant considerable attention to at least ensure we mitigate key risks, but ideally to also proactively capitalize on the opportunities.
Given how early and nascent the intersection of AI development and animal advocacy is, I believe that we need to put more effort into figuring out how to navigate these high stakes. Currently, many tangible and immediately actionable interventions can feel a bit unambitious and many ambitious interventions can feel intangible and riddled with uncertainty. What’s more, the complex nature of the stakes at play can, at times, feel genuinely paralyzing, making any attempts at meaningful research or field building at this intersection itself feel elusive.
In this post, I want to outline a couple of models and frameworks I have found myself internalizing that I believe are robust anchors, that are things that we do know when we approach the intersection of AI and animal advocacy. They are not mutually exclusive or collectively exhaustive, nor do they cover a consistent layer of granularity. But I hope they can help anchor people’s thoughts and serve as guiding mental models. Ideally, they can serve as nudges and starting points to develop compelling research questions and a strategic research agenda.
Influence and Values
Fundamentally, we want to increase pro-animal values in places of influence and to increase the influence of the pro-animal movement. In the context of AI, this brings about two key areas to look into:
- Getting animal advocates up to speed on AI. Understanding and adopting AI tools should, at a minimum, help us be generally more productive in our work, e.g., by streamlining research, drafting communications, or analyzing data. More indirectly, it may cultivate strategic consideration around AI-driven risks and opportunities for advocacy interventions. This understanding will hopefully also position the pro-animal movement more effectively as we continue to figure out the most promising and important implications of AI development.
- Influencing key AI stakeholders to become more pro-animal: In our work, we want to target the stakeholders with the most leverage. If we expect AI to have a huge impact on society, this may uncover new stakeholders (e.g., influential individuals within leading AI labs and companies) and raise the relative salience and importance of existing ones (e.g., governmental bodies in the US and China, given their outsized relative role in global AI development).
Some strategic questions anchored on this point then are:
Interest and Alternatives
Fundamentally, as a pro-animal movement, we want to increase interest in alternatives to factory farming (or decrease interest in factory farming itself) and ensure such alternatives exist. It seems unlikely that any key decision-makers hold an intrinsic interest in maintaining factory farming. Rather, insofar as factory farming is instrumental—to seek money or influence, feed the world, or provide sustenance and taste pleasure—there is, at least in theory, an alternative option. In the context of AI, this brings about three key areas to look into:
- Leveraging AI to improve alternatives. In a world massively transformed by AI, new alternatives may quickly arise, and existing alternatives may quickly improve. For consumers, this could include AI-driven welfare improvements on farms, AI-driven alternative protein R&D, or even, more speculatively, AI-driven alternatives to food entirely. Beyond consumers, this also includes alternative ways of making money, seeking power and influence, or feeding the world, addressing whichever interests various stakeholders in current factory farming systems have.
- Leveraging AI to increase interest in alternatives or decrease interest in factory farming. A lot of pro-animal advocacy is already focused on influencing stakeholder interest, such as consumer behavior. AI offers powerful new tools and strategic insights to enhance the effectiveness of these efforts, ranging from more targeted messaging to understanding and overcoming barriers to alternative adoption.
- Clearing obstacles for alternatives to be implemented. There are already many political, physical, and societal obstacles holding back the adoption of alternatives, and we may see many new obstacles in the near future. In the context of alternative proteins, work to address these obstacles may include preventing bans on cultivated meat, investing in the physical infrastructure necessary to scale up cultivated meat production and conducting research on (or developing) consumer interest in cultivated meat. A similar line of priorities may apply to other alternatives.
Some strategic questions anchored on this point then are:
- What obstacles may we face in implementing alternatives, and how can we use AI to overcome them?
- What are the intrinsic interests of different stakeholders? How can AI help identify and facilitate their transition away from factory farming?
- Which stakeholders have shared interests in which alternatives, and how do we get them on board? This would include allies such as environmentalists, who may seek a more efficient and sustainable alternative, as well as the industry itself, through its interest in an ultimately more cost-efficient way of producing similar products.
- How can AI facilitate demand-side interventions—beyond traditional advocacy—to shift cultural norms, dietary habits, and consumer preferences away from animal products at scale?
Symmetries and Asymmetries
Fundamentally, AI’s change on society will go far beyond the pro-animal movement. As we face large-scale societal transformation, it seems naive to assume that the pro-animal movement is the only one looking to leverage AI to improve its work. It is important to keep in mind how the industry may leverage the very same AI applications, often with competing or amplified effects. As such, it becomes strategically important to explore comparative advantages and ways in which animal advocates may deploy AI interventions to seek distinct strategic benefits. This consideration has been outlined in another post here.
Conclusion
The intersection of AI and animal advocacy is high stakes, but very complex and uncertain. By outlining these robust anchor points, I hope not to offer definitive answers, but rather to provide a clearer compass for navigation. These frameworks, while not exhaustive or perfectly granular, have been helpful for me to cut through the ambiguity of this space. I hope for them to act as nudges and starting points, empowering us to develop compelling research questions and a cohesive strategic research agenda that ensures that AI benefits animal welfare. Feel free to reach out if you have further ideas about anchor points and corresponding research questions. I would love to hear them. Some more things you can do - expect this to be my default call to action nowadays:
- Join the conversations on the Hive and the AI for Animals Slack spaces.
- Stay up to date with the AIA Newsletter and get caught up on key resources.
- If you work in animal advocacy, try to integrate AI into your work and stay on top of AI development. Some great people in our movement are writing about AI tools and offering support and advice.
- If you have questions or ideas about the intersection of AI and animals, write about them! Part of figuring out how we should handle these stakes is writing about them - and the Forum is not as intimidating a place as you might think :) And for what it’s worth, I’ll gladly offer another set of eyes for ideas!
Great post! I myself work at the intersection of AIxAnimals because I know it's important, but I feel like the lines in the Intro summarize it very well for me and I haven't quite put it that way in my mind before.