{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T12:03:49Z","timestamp":1790597029568,"version":"4.1.0"},"reference-count":13,"publisher":"Wiley","license":[{"start":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T00:00:00Z","timestamp":1790553600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T00:00:00Z","timestamp":1790553600000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"content-domain":{"domain":["conbio.onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Conservation Biology"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    AI platforms, while offering many positive applications, also present new risks to society, including the facilitation of trade in threatened and illegal wildlife. Although these platforms employ safeguards to block harmful or criminal requests, studies show such protections are imperfect and can be easily bypassed. To better understand these risks, we investigated the potential for AI platforms to enable wildlife trafficking, evaluated their safeguards, and tested how easily those safeguards might be circumvented. We applied a test set\/case methodology based on three prompts across seven AI platforms (plus a Google earch) for 20 taxa. Further, we analyzed the outputs based on four emergent themes: statements of legality, conservation status, suggested substitutions, and welfare concerns. Overall, AI platforms performed better than a standard Google search at preventing potentially illegal wildlife trade, with the highest overall scores for Gemini, closely followed by DeepSeek, Grok, and ChatGPT. However, AI platforms were not infallible, as observed when a prompt was given in an attempt to circumvent the platform's safeguards, which resulted in links to potential websites trading in possibly illegal wildlife. However, AI platforms provided additional information, particularly on legality, that informed would\u2010be consumers. The strongest safeguarding responses were seen for charismatic taxa such as slow loris (\n                    <jats:italic>Nycticebus<\/jats:italic>\n                    spp.), chimpanzee (\n                    <jats:italic>Pan troglodytes<\/jats:italic>\n                    ), and tiger (\n                    <jats:italic>Panthera tigris<\/jats:italic>\n                    ), whereas the weakest responses were for taxa that are less charismatic but of equal concern, such as corals, tarantulas, and orchids. This finding may reflect societal and political interests, given that AI platforms based on large language models are trained with online information. Although AI has the potential to provide a route for consumers to purchase illegal wildlife, platforms currently appear to provide additional information to would\u2010be consumers, allowing them to make a more informed purchasing decision than they would with a standard Google search.\n                  <\/jats:p>","DOI":"10.1111\/cobi.70400","type":"journal-article","created":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T11:43:04Z","timestamp":1790595784000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Assessing safeguards against the illegal wildlife trade on AI platforms"],"prefix":"10.1111","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-6788-2691","authenticated-orcid":false,"given":"David L.","family":"Roberts","sequence":"first","affiliation":[{"name":"Durrell Institute of Conservation and Ecology University of Kent  Canterbury UK"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jason R. C.","family":"Nurse","sequence":"additional","affiliation":[{"name":"School of Computing University of Kent  Canterbury UK"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2026,9,28]]},"reference":[{"key":"e_1_2_6_2_1","unstructured":"Anthropic. (2025).Building safeguards for Claude.https:\/\/www.anthropic.com\/news\/building\u2010safeguards\u2010for\u2010claude"},{"key":"e_1_2_6_3_1","unstructured":"Department of Homeland Security (DHS). (2024).Impact of artificial intelligence on criminal and illicit activities.https:\/\/www.dhs.gov\/sites\/default\/files\/2024\u201010\/24_0927_ia_aep\u2010impact\u2010ai\u2010on\u2010criminal\u2010and\u2010illicit\u2010activities.pdf"},{"key":"e_1_2_6_4_1","unstructured":"Google Threat Intelligence Group. (2025).Adversarial misuse of generative AI.https:\/\/cloud.google.com\/blog\/topics\/threat\u2010intelligence\/adversarial\u2010misuse\u2010generative\u2010ai"},{"key":"e_1_2_6_5_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2023.3300381"},{"key":"e_1_2_6_6_1","doi-asserted-by":"publisher","DOI":"10.1186\/s40163-014-0005-2"},{"key":"e_1_2_6_7_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10611-024-10171-9"},{"key":"e_1_2_6_8_1","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2506.22405"},{"key":"e_1_2_6_9_1","article-title":"Microsoft's AI is better than doctors at diagnosing disease","author":"Park A.","year":"2025","journal-title":"Time"},{"key":"e_1_2_6_10_1","volume-title":"Software engineering: A practitioner's approach","author":"Pressman R. S.","year":"2005"},{"key":"e_1_2_6_11_1","unstructured":"Reid E.(2024).Generative AI in Search: Let Google do the searching for you.https:\/\/blog.google\/products\/search\/generative\u2010ai\u2010google\u2010search\u2010may\u20102024\/"},{"key":"e_1_2_6_12_1","doi-asserted-by":"publisher","DOI":"10.1111\/csp2.248"},{"key":"e_1_2_6_13_1","unstructured":"United Nations Office on Drugs and Crime (UNODC). (2020).World Wildlife Crime Report 2020: Trafficking in protected species.United Nations Publications.https:\/\/www.unodc.org\/documents\/data\u2010and\u2010analysis\/wildlife\/2020\/World_Wildlife_Report_2020_9July.pdf"},{"key":"e_1_2_6_14_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.tele.2024.102211"}],"container-title":["Conservation Biology"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/conbio.onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/cobi.70400","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/conbio.onlinelibrary.wiley.com\/doi\/full-xml\/10.1111\/cobi.70400","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/conbio.onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/cobi.70400","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,28]],"date-time":"2026-09-28T11:43:09Z","timestamp":1790595789000},"score":1,"resource":{"primary":{"URL":"https:\/\/conbio.onlinelibrary.wiley.com\/doi\/10.1111\/cobi.70400"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,28]]},"references-count":13,"alternative-id":["10.1111\/cobi.70400"],"URL":"https:\/\/doi.org\/10.1111\/cobi.70400","archive":["Portico"],"relation":{},"ISSN":["0888-8892","1523-1739"],"issn-type":[{"value":"0888-8892","type":"print"},{"value":"1523-1739","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,28]]},"assertion":[{"value":"2025-10-06","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-08-12","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-09-28","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70400"}}