Disturbing Bullfighting Study Has Advocates Raising the Alarm Over AI Travel Agents

A newly posted research preprint has dropped a bombshell on the travel industry: AI travel agents will schedule a bullfight for you without blinking — unless someone explicitly tells them not to.

Hollywood, California, June 29th, 2026 – The fascinating study, titled “Your AI Travel Agent Would Book You a Bullfight,” introduced a benchmark called TAC — Travel Agent Compassion — and the results were not pretty. Researchers found that AI travel agent systems routinely failed to steer users away from animal-exploitation tourism unless animal welfare was written directly into the system prompt. This is a preprint, not law, but the implications are impossible to ignore.

AI & Animals: A Short Documentary

The TAC benchmark tested a simple question: would an AI travel agent avoid tourism options built on animal suffering? The answer, largely, was no. Left without explicit guidance, the AI treated a bullfight the same as a museum ticket — no hesitation, no flag, no warning. Add animal-welfare language to the prompt, and the system suddenly became far more discerning. The difference between a compassionate AI and a harmful one, the researchers found, came down to a few lines of instructions.

That gap is a serious problem. AI adoption across the tourism industry has expanded rapidly, with chatbots and recommendation models now standard across major booking platforms, according to a Frontiers research review on AI in tourism. When these systems optimize for what is popular, cheap, or highly rated, they surface a wide menu of attractions that animal advocates consider outright exploitation — bullfighting and blood spectacles framed as cultural events, marine mammal shows featuring captive dolphins and orcas, elephant rides and trekking camps, animal performances in circuses and roadside attractions, and captive wildlife selfie setups. None of this is flagged. None of it is questioned.

The deeper issue is who is teaching these systems what animals are worth. An AI travel agent does not hold values — it absorbs them from training data, business incentives, and the priorities of whoever builds it. If that data treats animals as commodities, the AI will too, quietly and at a scale that dwarfs any individual booking agent. Plant-based interest is climbing fast, with events like the Great Vegan Gathering festival in London drawing record crowds, yet the machines booking our vacations have not registered the shift.

UnchainedTV‘s documentary “AI & Animals: A Short Doc” examines exactly this fork in the road — whether AI entrenches animal exploitation or becomes a force for compassion. The film’s central argument tracks directly with the study’s findings: the future is being coded right now, and animals are at serious risk of being left out of the equation entirely. The channel’s explainer on whether AI will harm or help animals, maps both directions this technology could take.

Aligned with profit and convenience, AI could automate harm at industrial scale. An AI travel agent chasing cheap, popular, high-rated bookings will keep filling seats at animal shows. The same logic that applies to tourism applies to industrial agriculture, animal testing pipelines, and entertainment systems built on captive animals. When a trusted digital assistant books a marine park visit without comment, it normalizes the activity for millions of users simultaneously. Concerns raised in coverage of the fight against dog testing show how entrenched these systems already are — and how hard they are to dismantle once embedded.

The same technology, however, can be pointed in the opposite direction. Built with animal welfare as a design criterion, AI can promote non-animal research methods, advance plant-based and cultivated protein systems, expose cruelty in supply chains, and improve public understanding of animal rights issues.

How You Can Help Animals When You’re Using AI

Every animal lover who uses AI needs to give it a literal thumbs up when it produces pro-animal, pro-compassion content and give it a thumbs down when it treats animals as commodities. The option is right there to click at the bottom of every bit of content AI produces. Additionally, when a user sees AI producing recommendations, bookings or any content that promotes animal exploitation, animal lovers are urged to click on the gear icon and follow the prompts to lodge a complaint. This is one way AI can learn to respect animals.

Researchers are already working on decoding animal communication. An AI travel agent trained on welfare criteria could route travelers toward genuine sanctuaries rather than exploitation attractions. The study makes the fix sound almost straightforward: tell the system to care, and it largely does. The harder work is making that the default, not an optional add-on. The growing interest in compassion-driven living — from faith-based perspectives on animals to everyday consumer choices — shows the cultural appetite is already there.

Consumer reliance on AI to plan travel, manage schedules, and make recommendations is accelerating. As that reliance deepens, so do the stakes. Animal advocates argue that welfare safeguards must be built into these systems now, before automated exploitation becomes the invisible, unquestioned standard. A compassionate AI travel agent will not emerge by accident. It will exist only if developers are pressured to build one, users demand it, and animal protection is treated as a non-negotiable design requirement — not an afterthought left for someone else to add later.