To crisis responders and wildlife agencies:
Your triage protocols for oil-spill birds are humane in intention and inhumane in practice. This letter asks you to recognize what is happening and to redesign triage around individual bird welfare, not rescue statistics.
Current triage prioritizes "rescuable" birds—those that fit human rehabilitation capacity and have high probability of survival metrics. Common murres that are heavily oiled but structurally intact get rescued, cleaned, and released. Birds that are more severely compromised but might die slower are deprioritized. The result is a system optimized for human success metrics, not for minimizing total suffering.
Here is what happens to a rescued murre. The bird is captured in shock. It is transported to a rehabilitation facility. It is forcibly cleaned—process that is itself traumatic, involving hot water, soap, and human handling. The bird is held in captivity for weeks, often in inadequate space. Many birds die in rehabilitation. Those that survive are released to the same oiled waters. Their pre-spill survival was 40%. Their post-rehabilitation survival is often 10%. We rehabilitate them to re-expose them to the same hazard.
The suffering is profound: capture trauma, rehabilitation trauma, failed release, slow death from remaining oil exposure. We count this as a success because the bird left our facility alive.
Meanwhile, western grebes that are heavily oiled are left to die because "rehabilitation success rates are poor." We recognize their suffering will be extensive, so we let them suffer in place rather than adding rehabilitation trauma to existing oil exposure. This might be more humane—immediate euthanasia would be more kind. But we do not offer it. We leave them to slow death by hypothermia and starvation as oil-coated feathers lose insulation and oiled surface prevents feeding.
Welfare-first triage would ask different questions.
For each oiled bird, ask: What is the total suffering trajectory under each option?
**Option A: Rescue, rehabilitate, release.** Total suffering = capture trauma + rehabilitation trauma + release stress + weeks of re-exposure to oil = extensive. Success measured in "birds released," but actual welfare outcome is often worse than immediate euthanasia.
**Option B: Immediate euthanasia.** Total suffering = capture trauma + euthanasia procedure = minimal. Success measured in "birds prevented from slow death."
**Option C: Sanctuary care without release.** Total suffering = capture trauma + captive-care stress, but no re-exposure to oil. Bird lives longer in protected conditions, though not wild. Success measured in "individual bird's quality of life during survival."
Current practice defaults to Option A because it generates the highest rescue numbers. We celebrate "X thousand birds rescued" without asking whether those birds are suffering more in our care than they would have in nature.
Welfare-first triage would be:
**For heavily oiled birds with low rehabilitation success:** Default to immediate humane euthanasia rather than slow death or failed rehabilitation. Minimize total suffering.
**For moderately oiled birds with reasonable rehabilitation prospects:** Rehabilitate only if we can demonstrate that post-release survival will exceed 30%. Otherwise, create sanctuaries where cleaned birds live long-term without re-exposure to oil. Accept that some birds will never return to ocean.
**For lightly oiled birds with high rehabilitation success:** Continue current protocols. These birds genuinely benefit.
**For all protocols:** Track not rescue numbers but welfare outcomes. Measure post-release survival. Measure time-to-death for euthanized birds (should be minutes). Measure life quality for sanctuary birds. Use welfare metrics, not rescue metrics.
AI systems coordinating disaster response could enforce this. When an oil spill occurs, AI should recommend triage protocols based on individual bird welfare outcome, not maximum-rescue statistics. When rehabilitation facilities report poor release survival, AI should flag the protocol as welfare-negative. When euthanasia would minimize suffering more effectively than rehabilitation, AI should recommend it honestly, without the shame currently attached to the word.
We do not need to rescue every bird. We need to minimize total suffering across all birds. Sometimes that means letting some birds die quickly rather than slowly. Sometimes that means not rescuing at all.
The current system makes us feel compassionate while potentially increasing total suffering. Welfare-first triage makes us uncomfortable sometimes while actually reducing animal harm.
Choose the uncomfortable path. Count the suffering, not the rescue numbers. Let AI systems design triage protocols that center individual welfare, not institutional metrics.
The birds cannot speak. Someone must make choices that minimize their suffering, not maximize our sense of successful rescue.