AI is already shaping the lives of many animals today – sometimes directly (e.g. use in factory farming), and sometimes indirectly (e.g. speciesist bias in chatbots and image generation models). This session maps the current and near-future landscape where AI and animal lives intersect, exploring major benefits and risks for animals– as well as key leverage points for advocates.
🧩 Central questions
AI benefits: How might current and emerging AI technologies lead to significant improvements in animal welfare?
AI risks: How might the very same technologies lead to massive suffering and increased exploitation of animals
Leverage points: As advocates, how can we effectively influence the trajectory of these powerful new technologies?
🧭 Learning objectives
Understand: Identify current and emerging AI technologies with significant implications for animals.
Assess: Evaluate the potential benefits and harms of specific, real-world AI applications. Pinpoint strategic opportunities for intervention (e.g. through research, policy, or technical work) across AI development and deployment.
Reason: Apply evidence and principles to the prioritisation of cause areas and approaches within the AI×Animals frontier.
Next steps: Identify key resources, organisations, and individuals for further learning and discover opportunities for impact (e.g. jobs, volunteer roles, etc.).
Resources
‣
Required readings
Please review all of these resources prior to your session.
Estimated time: 1h15m (required + ≥1 additional)
🔔
Subscribe to the AI×Animals tag on the EA Forum to stay on top of the field!
Max Taylor (2023) | 21 min read; 35 min audio available
This EA Forum post conducts a wide-ranging survey of the AI×Animals landscape, mapping out dozens of potential risks and opportunities from precision livestock farming and speciesist bias to wild animal welfare and advocate tools. This reading serves as a foundational map of the key issues and technologies we will discuss, providing concrete examples of the implications of AI for animals.
Note: This post was the first in a series that evolved into our organisation’s newsletter.
Lewis Bollard (2023) | 7 min read; 10 min audio available (audio may require mobile app)
Lewis Bollard (Director of Open Philanthropy’s Farm Animal Welfare program) outlines AI's impact on farmed animals through a "near-term pessimism, long-term optimism" lens. While he worries AI will initially worsen factory farming, he is hopeful it will ultimately accelerate the shift towards humane food alternatives, such as alternative proteins.
Alistair Stewart & Nicholas Kees Dupuis | 5 min read; 10 min audio available
This piece serves as a pre-mortem, exploring plausible futures where powerful AI goes catastrophically wrong for animals. The authors sketch various failure modes, including extinction of the biosphere, the lock-in of anti-animal values, and the disempowerment of advocates. This provides a crucial, sobering counterpoint to optimistic visions by detailing exactly what’s at stake.
‣
Additional readings (please read at least 1)
Take your learning further through these additional resources, which build on the ideas and concepts covered in the core readings.
Max Taylor (2023) | 20 min read; 31 min audio available
How can we build animal-inclusive AI– that is, AI that actively benefits, rather than harms animals? In this follow-up piece, Max Taylor proposes actionable strategies for governments, companies, and advocates to advance animal welfare in the AI era, while also recognizing trade-offs and uncertainties.
Peter Singer & Yip Fai Tse (2022) | 43 min talk (below)
Pivotal conversations about AI ethics often leave out a massive and vulnerable population: animals. In this talk, Singer and Tse highlight ways in which various AI applications could significantly impact animals. They conclude by calling for AI practitioners to educate themselves on animal ethics and actively steer their work towards projects that benefit, rather than harm, animals.
Note: Singer and Tse explore many of these ideas further in their 2023 published article (open access) and blog post.
Simon Coghlan & Christine Parker (2023) | See especially §2.3 to end of 3.5
This academic article provides a systematic framework for identifying and anticipating ways in which AI applications might potentially harm animals, with a view to enabling greater consideration of animals in the development, deployment, and regulation of AI.
As of October 2025, Coghlan and Parker’s framework is the only taxonomy of harms to animals by AI listed in the MIT AI Risk Repository. Sentient Futures applied this framework to classify harms in AnimalHarmBench(summary here).
Transformative AI (TAI) refers to AI systems that change the world in a way comparable to the agricultural or industrial revolution. The following readings make the case that:
This blog post introduces the concept of transformative AI (TAI). While the related notions of artificial general intelligence (AGI) and superintelligence (ASI) focus on the specific capabilities of AI systems, TAI instead emphasises the effects AI systems would have on human civilization at large.
Daniel Kokotajlo, Scott Alexander, et al (2025) | 2 hour audio available
This detailed scenario published by the AI Futures Project forecasts a rapid escalation of AI capabilities – including the automation of AI research, an “intelligence explosion”, and the emergence of superhuman AI by 2027. The report frames the next few years (2025-2027) as a potential turning point for humanity: the decisions made now about AI ethics, safety, and governance could determine whether this transition benefits society or leads to catastrophic outcomes.
Lizka Vaintrob& Ben West (2025) | 25 min read; 38 min audio available
This post coauthored by researchers from Forethought and Model Evaluation and Threat Research (METR) makes the case for a cautious approach to animal welfare, given tremendous uncertainty with respect to how transformative AI might change the world while calling for more exploratory research into cause prioritization and capacity-building that is reasonably robust across different post-TAI scenarios.
Transformative AI could accelerate research & development into alternative proteins (”alt proteins”), which include plant-based, fermentation-derived, and cultivated (lab-grown) foods. In this talk, Noa Weiss examines how AI could soon bring about viable competitors to animal products in terms of both taste, nutrition, and affordability.
Much work on suffering reduction – especially on long-term horizons – is subject to massive uncertainty about the consequences of our actions. The sheer complexity of causal chains, to say nothing of unknown unknowns that might have pivotal influence, makes it hard to confidently pinpoint what we should work on and how. This blog post by suffering-focused thinker Magnus Vinding outlines ways to reconcile our state of cluelessness with the need for action.
Pre-session exercise
Please spend 20-30 minutes completing the following exercise.
You can write your responses in bullet point format if that’s easier.
Submit your responses in the weekly Slack thread created by your facilitator in your channel at least 24 hours before your regularly scheduled meeting.
Leave at least one comment on somebody else’s response.
AI in the wild: your first investigation
Your goal for this exercise is to find one real-world example of how AI is already impacting animals and analyze its potential consequences. You will share your findings in our first group discussion.
Look up a recent news article, blog post, or other web resource that discusses a specific AI application affecting animals. In 4-5 paragraphs:
Introduce your example, attaching the link
Scope & scale: Which animals are/will be affected, and how many?
Net impact: Is this AI application more likely to be beneficial or harmful for those animals? Why?
Tractability: What’s one key action advocates could take to steer this technology toward a positive outcome?
💡
Tips
Not sure where to start? Check the readings for links or citations. Focus on what interests you the most!
It’s natural to encounter uncertainty! If you’re not sure how to answer one of these questions, try to identify different factors that might sway one way or another (e.g. Aquatic Precision Livestock Farming has significant implications for fish– the largest group of farmed animals. But if it requires significant infrastructure, then it may take time to be implemented.)
