Thanks for this very clear wake-up call. I’ve been wrestling with similar questions. I think this presents measured and sensible approaches.
I fully agree with your “This matters for animals and their advocates” section. The idea that AI issues are just for tech bros is such a dangerous blind spot. The potential scenarios you outline (intensified factory farming, ecosystem lock-in, breakthrough technologies) really drive home how transformative this could be for animals specifically.
I also strongly agree with the “build capacity” approach, especially your point that theories of change themselves need to shift. I think you’re right that many current AI-for-animals efforts focus on improving efficiency within existing frameworks rather than grappling with how fundamentally different the strategic landscape might become.
On AI welfare, the research-first approach you highlight makes sense, any efforts here need to be careful, measured, and credible given the risks of getting this wrong.
I’m less convinced that “optimise for immediate results” is the right response to uncertainty about AI trajectories. If we’re truly in a transition period, the most important work might actually be the hardest to measure in the short term, like influencing the foundational assumptions being built into emerging systems right now.
The “predict and prepare” approach feels most promising to me, but I think we might be too focused on specific AI applications (like cultivated meat acceleration) we’re likely heading into entirely new economic and governance systems where the basic rules about value, ownership, and moral consideration are being rewritten. The assumptions embedded during these transition periods could determine whether animals are treated as production units or moral patients for generations. Therefore the economic frameworks emerging alongside AI development could be just as consequential as the AI systems themselves.
Questions I’m asking myself, and feel free to posit any answers! Are animal advocates engaging enough with the economic/governance transitions happening alongside AI development or are we too narrowly focused on the technology itself? Are we thinking big enough about the window of opportunity during these transitions?
Thanks again for this piece, I think it’ll make a good shareable read to help others in our organisations understand the questions we should be asking/ what we should be doing.
Thanks Karen! Interested if you have specific things in mind for implications of the economic angle? I can certainly see it playing into some of the “Predict how AI will change things, and try to make that go well for animals” predictions, or leading to more of an emphasis on “Shift towards all-inclusive AI safety”.
Great question! I’m thinking about how the economic disruptions from AI create opportunities to reshape the foundational rules before new systems crystalize.
For example, as AI automates more labour and potentially destabilises growth-oriented models, we might see experiments with post-growth economics, universal basic services or entirely new frameworks for measuring value. These transition moments are when assumptions about what “counts” economically become malleable, including whether animals are seen as production inputs or beings with inherent worth.
Right now, our economic systems have deeply embedded assumptions that treat animals as commodities, externalise ecological costs and prioritise efficiency over welfare. But during systemic transitions, these assumptions become visible and potentially changeable in ways they normally aren’t.
I think this fits most naturally into your “predict and prepare” category, but with a focus on economic system design rather than just technological applications. Instead of just preparing for cheaper cultivated meat, we might also prepare for the governance frameworks that will determine how new economic models treat animals.
The policy levers might be things like: ensuring animal welfare is embedded in any new economic measurement systems, preventing harmful defaults from getting locked into emerging governance structures or influencing how post-growth economic experiments value different forms of life.
Does that distinction between technological applications and systemic economic design make sense? I suspect the latter might be more neglected right now.
I’ve been exploring some of these ideas in more depth [here].
Thanks for this very clear wake-up call. I’ve been wrestling with similar questions. I think this presents measured and sensible approaches.
I fully agree with your “This matters for animals and their advocates” section. The idea that AI issues are just for tech bros is such a dangerous blind spot. The potential scenarios you outline (intensified factory farming, ecosystem lock-in, breakthrough technologies) really drive home how transformative this could be for animals specifically.
I also strongly agree with the “build capacity” approach, especially your point that theories of change themselves need to shift. I think you’re right that many current AI-for-animals efforts focus on improving efficiency within existing frameworks rather than grappling with how fundamentally different the strategic landscape might become.
On AI welfare, the research-first approach you highlight makes sense, any efforts here need to be careful, measured, and credible given the risks of getting this wrong.
I’m less convinced that “optimise for immediate results” is the right response to uncertainty about AI trajectories. If we’re truly in a transition period, the most important work might actually be the hardest to measure in the short term, like influencing the foundational assumptions being built into emerging systems right now.
The “predict and prepare” approach feels most promising to me, but I think we might be too focused on specific AI applications (like cultivated meat acceleration) we’re likely heading into entirely new economic and governance systems where the basic rules about value, ownership, and moral consideration are being rewritten. The assumptions embedded during these transition periods could determine whether animals are treated as production units or moral patients for generations. Therefore the economic frameworks emerging alongside AI development could be just as consequential as the AI systems themselves.
Questions I’m asking myself, and feel free to posit any answers! Are animal advocates engaging enough with the economic/governance transitions happening alongside AI development or are we too narrowly focused on the technology itself? Are we thinking big enough about the window of opportunity during these transitions?
Thanks again for this piece, I think it’ll make a good shareable read to help others in our organisations understand the questions we should be asking/ what we should be doing.
Thanks Karen! Interested if you have specific things in mind for implications of the economic angle? I can certainly see it playing into some of the “Predict how AI will change things, and try to make that go well for animals” predictions, or leading to more of an emphasis on “Shift towards all-inclusive AI safety”.
Great question! I’m thinking about how the economic disruptions from AI create opportunities to reshape the foundational rules before new systems crystalize.
For example, as AI automates more labour and potentially destabilises growth-oriented models, we might see experiments with post-growth economics, universal basic services or entirely new frameworks for measuring value. These transition moments are when assumptions about what “counts” economically become malleable, including whether animals are seen as production inputs or beings with inherent worth.
Right now, our economic systems have deeply embedded assumptions that treat animals as commodities, externalise ecological costs and prioritise efficiency over welfare. But during systemic transitions, these assumptions become visible and potentially changeable in ways they normally aren’t.
I think this fits most naturally into your “predict and prepare” category, but with a focus on economic system design rather than just technological applications. Instead of just preparing for cheaper cultivated meat, we might also prepare for the governance frameworks that will determine how new economic models treat animals.
The policy levers might be things like: ensuring animal welfare is embedded in any new economic measurement systems, preventing harmful defaults from getting locked into emerging governance structures or influencing how post-growth economic experiments value different forms of life.
Does that distinction between technological applications and systemic economic design make sense? I suspect the latter might be more neglected right now.
I’ve been exploring some of these ideas in more depth [here].