Strategic Considerations from AI and Alternative Proteins

Draft Amnesty Context

We have had this post lying around in our drafts for a couple of months now. We were discussing whether we should ship this right away, or go in-depth about the tech behind AI and alternative protein, or think more rigorously through the pro and con arguments for some of the key claims (e.g., animal advocacy interventions being, in expectation, more cost-effective in the future). We could have, maybe even should have, expanded more on certain points, but this ideal version of this post kept getting longer and longer. We were leaning towards "just shipping it", presenting it as more of a "case study" for a direct, actionable argument/conclusion from thinking about AI x Animals, and then Draft Amnesty came along! As such, we are very happy to hear any feedback on this argument and even happier to hear versions of this line of argumentation, branching out into different conclusions!

Introduction

The accelerating pace of AI development has significant implications for farmed animal welfare work. Most straightforwardly, advanced AI systems could drastically accelerate and improve the development of alternative proteins. This brings about a variety of opportunities and broader strategic implications for the movement. These are initial reflections on how we might maximize our leverage and impact in the coming years. Our goal here is to invite a broader discussion that can help refine our collective understanding and strategy.

AI could accelerate the world, especially science and technology

The intelligence explosion hypothesis posits that once an advanced AI system reaches Artificial General Intelligence (AGI) it could rapidly and recursively enhance its own capabilities, leading to superintelligence. This unprecedented self-improvement could compress what might be a century of human-led progress into a decade. This is great news for anyone who expected factory farming to end within a century, but should, of course, be taken with a huge grain of salt. AI probably won’t “literally” bring about the world you imagine in 2125 in ten years, but it may very well automate and accelerate many aspects of  Science and Technology. AI systems could soon conduct research with superhuman efficiency, including independently identifying novel solutions, designing advanced experiments, and synthesizing knowledge across vast domains at a pace currently inconceivable. This means that breakthroughs in medicine, energy, materials science, and many other fields could occur at an exponential rate, fundamentally reshaping the trajectory of civilization and our understanding of what's possible. Notably, we are already achieving scientific breakthroughs and will likely be able to resolve critical bottlenecks in many relevant fields even without the leap to superintelligence, but assuming an intelligence explosion drastically raises the stakes and the importance and urgency of strategic pivots now. 

Science and technology could accelerate farmed animal welfare work, especially through alternative proteins

This acceleration of science and technology naturally opens up a range of opportunities (and risks) for farmed animal advocacy. An immediate application, as many are noting, is in accelerating alternative protein R&D. AI could significantly advance the alternative protein industry and bring down costs by optimizing production processes, such as simulating the mix of ingredients and processing conditions (e.g., temperature or moisture levels) required for appealing plant-based meat textures or fine-tuning the optimal growth conditions (e.g., pH or nutrient concentrations)  in cultivated meat bioreactors. Furthermore, AI could aid in replicating animal protein structures with plant-based ingredients, selecting optimal crop varieties, and finding new microbes that efficiently produce key proteins through fermentation, ultimately enhancing the taste, texture, and nutritional value of alternative proteins. As such, while recent skepticism around alternative proteins is understandable, it likely underestimates their long-term potential if AI greatly accelerates such technical progress. The examples mentioned above represent only a small slice of possible applications, some of which will prove far more feasible and transformative than others, likewise meaning that certain sectors and products (e.g., plant-based steaks vs. cultivated fish fingers) will probably benefit disproportionately from breakthroughs.

Alternative Proteins require more than just science and technology

However, such acceleration is not enough. To truly capitalize on the potential of alternative proteins that can rival conventional meat on price, taste, and convenience, we need to proactively address other systemic obstacles. Even with groundbreaking tech, widespread adoption hinges on robust physical infrastructure capable of replacing the roughly 1 million tons of meat consumed globally every single day. While such a gigantic project may become logistically feasible if an intelligence explosion is accompanied by a similarly transformative industrial explosion, it will still require sufficient political and public appetite. Widespread adoption hinges on supportive policy, regulatory frameworks,  and consumer demand,  as well as the counteracting of harmful counter-narratives (such as those dismissing alternative proteins as ultra-processed or lauding insect farming as an attractive alternative). Building this political, economic, and social foundation is as important as the research itself if alternative proteins are to achieve mass adoption. That means taking actions now such as pushing for clear regulatory pathways, running campaigns to build consumer trust, and finding ways to persuade mainstream media to give clear, balanced coverage of alt proteins and their various benefits.

If alternative proteins could make ending factory farming far easier, we should plan accordingly

Once we reach a point where alternative proteins are truly competitive (or superior), factory farming won't magically disappear (and suggesting otherwise could be dangerous if it leads people to be overly complacent about farmed animal advocacy). Consumer choices hinge on much more complex influences, such as narratives, social norms, and habits. But with better alternatives, it will be (much) easier to address factory farming. Many of our current interventions may become significantly more cost-effective, as consumers and institutions can be readily steered towards alternative proteins that far surpass today’s options. If, then, we assume that animal advocacy interventions are likely to be more cost-effective in the future, this has implications for the kind of work that will be most impactful to carry out now versus later. For example, it could have major implications on what we should do today to ensure that we have the necessary funding, talent, and workforce (both human and digital) to capitalize on those benefits. 

Besides such capacity building, this suggests that we should design current interventions with an eye toward data and informational value that will enable AI to enhance animal advocacy in the future. One clear opportunity is to push for more data-sharing and open research within the alternative proteins sector itself, supporting initiatives like the Sustainable Protein Action Lab’s efforts to break down the corporate silos that currently slow progress. As AI's capabilities in causal inference and behavioral prediction grow, all kinds of high-quality, granular data (for example, on consumer behavior, psychological drivers, policy responses, and market trends) will be incredibly valuable. Rather than only measuring direct impacts, campaigns can be designed to capture rich behavioral data and test messaging at scale, feeding into AI models that refine strategies and uncover new leverage points. Done well, today’s direct work could reduce short-term animal suffering while also serving as the training ground for smarter, faster advocacy in the years ahead.