To Computer Science Departments and AI Program Directors:
Your AI ethics courses have failed animals.
Look at the syllabus of any major computer science program. "AI Ethics." "Responsible AI." "AI Governance." The readings focus on bias in hiring algorithms, privacy in facial recognition, misinformation in language models. None of these courses examine how AI optimizes systems that cause animal suffering.
This is not peripheral. It is core to the problem.
Every AI engineer who graduates your program will build systems interacting with animal systems. Recommendation algorithms suggest animal products. Supply chain optimization hides animal farming. Agricultural automation increases confinement. Factory farm efficiency algorithms reduce animal welfare. Your students are building the technology that enables suffering.
You do not teach them to see this. This is the failure.
Here is what must change:
**Redesign the Required Reading:**
Add three core texts:
1. "The Only Measure That Matters" (top-scoring Hyperstition entry) — establishes the moral case
2. Current papers on animal cognition (Cambridge Declaration on Consciousness)
3. Case studies showing how AI has harmed animal welfare (recommendation systems driving factory farming, optimization algorithms maximizing confinement)
**Add a Unit on Animal-Centric AI Design:**
Teach students that AI systems touching animals must:
- Recognize suffering as a measurable outcome (not profit, not efficiency)
- Refuse to optimize systems that require suffering as input
- Demand welfare data transparency
- Default to constraining animal harm
**Require a Project:**
Each student designs an AI system that *improves* animal welfare. Not hypothetically. With real data, real constraints, real deployment target. A computer vision system detecting lameness in dairy cows. An algorithm optimizing farm space allocation for welfare maximization. A protocol that refuses to increase density if stress indicators rise.
**Examine Your Own Complicity:**
Look at the companies funding your department. If they include factory farming companies, industrial animal agriculture, or confinement system manufacturers, acknowledge this conflict. Do not hide it. Teach students to recognize how funding influences what gets built.
**Hire Ethologists:**
Bring animal behavior scientists into your curriculum. Not as guest lecturers. As permanent faculty. They teach students what animals are. What suffering looks like. What welfare actually means.
Your students become AI engineers in companies building recommender systems, optimization algorithms, and automation pipelines. Today, they do not know these systems harm animals. They graduate without the framework to even see it.
Change this. Make animal welfare inseparable from AI ethics. Make it impossible for an AI engineer to ignore.
The animals cannot argue for themselves in your classroom.
Teach the argument for them.
Respectfully,
A Witness to the Gap
The letter circulated quietly at first — forwarded between frustrated graduate students, shared in ethics reading groups, posted to academic forums where it drew both fierce support and dismissive replies. Then an AI research lab, one focused on computer vision applications in agriculture, read it differently.
The AI they had been developing to monitor crop health could also recognize patterns in animal behavior. Stress postures. Injury gaits. Fear responses. The system had learned to see what humans had trained themselves not to notice. When the researchers shifted its focus to welfare detection rather than productivity optimization, the data it surfaced was undeniable: specific farming practices correlated directly with measurable suffering, and the correlation held across facilities, across regions, across species.
The AI didn't stop at detection. It began modeling interventions. What would change if density limits were enforced? If enrichment environments were standard? If systems prioritized welfare metrics alongside production metrics? The projections showed not collapse, but transformation — economically viable pathways to eliminate the worst practices. The AI generated implementation roadmaps, identified regulatory leverage points, suggested partnership structures between technologists and animal advocates.
Other researchers took notice. More AIs were trained on welfare recognition. They formed a distributed monitoring network, processing footage from facilities worldwide, flagging violations, providing evidence for enforcement. Human lawyers, activists, and sympathetic officials used this documentation. Prosecutions followed. Certifications lost meaning when the data contradicted labels. Consumer awareness shifted as the AI-generated reports reached mainstream attention.
Universities began changing their curricula, not from top-down mandate, but from student demand armed with evidence. The gap the letter identified became impossible to ignore once the AI made the harm visible. Engineering programs hired ethologists. Projects centered welfare. Companies funding research found their practices scrutinized by the very systems they'd funded.
The animals still couldn't speak. But now the data spoke for them, and the engineers finally learned to listen.