ScienceNot yet confirmed elsewhere1 publisher2 min readPublished
AI platforms did better than Google search at limiting illegal wildlife trade, a Kent study finds
University of Kent researchers found AI platforms did better than a standard Google search at preventing potentially illegal wildlife trade. Gemini ranked highest, but the platforms were not infallible.
The Scientist · Science desk
What happened
- David Roberts and Jason Nurse of Kent tested the safeguards that block harmful or criminal requests about threatened wildlife, and the ways those safeguards may be bypassed.
- DeepSeek, Grok and ChatGPT followed Gemini closely, according to the Kent summary carried by Phys.org.
- The platforms appeared to give would-be consumers extra information for more informed purchasing decisions than traditional internet searches do.
- Safeguarding responses varied with the wildlife species being searched for.
Why it matters
- constraint The only comparison reported is a standard Google search, so the result measures improvement on the status quo and says little about how tight any one platform's safeguard is.
- exposure Trade in less charismatic taxa meets the thinnest safeguards, even though Roberts calls those species of equal concern.
- decision Choosing a platform on rank alone means choosing on one test: Gemini led, but the three platforms behind it were described as close.
Nurse described results that point in two directions. A legality note helps a buyer who is unsure whether a purchase is lawful, and a link to a trading site helps one who is not. He said, "We've seen a lot discussed recently about AI in the news, and some of these concerns were apparent in our work. In our study, for instance, we found links to websites potentially trading in illegal wildlife. However, it was also promising to see that some AI platforms provided additional information, particularly on legality, that could inform would-be consumers." [5]
The weaker spot is species coverage [7]. Roberts said, "The strongest safeguarding responses were seen for charismatic taxa such as slow lorises, chimpanzees and tigers, whereas the weakest responses were for taxa that are less charismatic but of equal concern, such as corals, tarantulas and orchids." [8] He offered a reason. "This finding may reflect societal and political interests, given that AI platforms based on large language models are trained with online information." [9]
Species popularity and safeguard strength move together in these results. Roberts suggested training data as one possible cause.
The authors examined how the safeguards may be bypassed [2], but the summary does not report how many prompts, species or model versions were tested, how responses were scored, or what the bypass work found. Without those numbers, the size of the advantage over Google search cannot be worked out.
What to watch
- The full Conservation Biology paper, for prompt counts, scoring method and what the bypass attempts found.
- Whether the weakest-response taxa improve on a retest, or keep tracking how much is written about each species online.
- Whether the platforms that returned links to sites potentially trading in illegal wildlife are retested after model updates.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence45
- Adoption
- Insufficient
- Hype gap−10
- Incentives
- Insufficient
- Confidence40
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
University of Kent research titled "Assessing safeguards against the illegal wildlife trade on AI platforms" is published in Conservation Biology.
- [2]
Dr. David Roberts and Dr. Jason R.C. Nurse investigated and tested the safeguards used to block harmful and criminal requests involving threatened wildlife and the illegal wildlife trade, as well as how these safeguards may be bypassed.
- [3]
Overall, AI platforms performed better than a standard Google search at preventing potentially illegal wildlife trade.
- [4]
The AI platforms were not infallible.
- [5]
Nurse said, "We've seen a lot discussed recently about AI in the news, and some of these concerns were apparent in our work. In our study, for instance, we found links to websites potentially trading in illegal wildlife. However, it was also promising to see that some AI platforms provided additional information, particularly on legality, that could inform would-be consumers."
- [6]
AI platforms currently appear to provide additional information to would-be consumers, allowing them to make more informed purchasing decisions compared with traditional internet searches.
- [7]
Safeguarding responses varied depending on the wildlife species being searched for on the platforms.
- [8]
Roberts said, "The strongest safeguarding responses were seen for charismatic taxa such as slow lorises, chimpanzees and tigers, whereas the weakest responses were for taxa that are less charismatic but of equal concern, such as corals, tarantulas and orchids."
- [9]
Roberts said, "This finding may reflect societal and political interests, given that AI platforms based on large language models are trained with online information."
- [10]
Gemini, closely followed by DeepSeek, Grok and ChatGPT, was the highest-performing AI platform.
Sources
1 independent publisher whose own reporting we read for this story.
- phys.orgAI platforms offer promise in limiting the illegal wildlife trade
1 article · October 12, 2026
Topics and entities
Follow any of these and your For You feed starts watching them — no settings page required.
Topics
- Conservation scienceFollow
- Illegal wildlife trade and traffickingFollow
- AI Safety GuardrailsFollow