Citizen science biodiversity platforms are confronting a growing vulnerability as generative artificial intelligence enables the rapid proliferation of fabricated species records. Independent entomologist Kris Anderson recently documented a synthetic dead leaf mantis submitted to iNaturalist, raising alarms about the reliability of user-generated ecological data. The incident follows similar discoveries, including a heavily altered avian record of a willow tit in Scotland, examined by Alexander Lees, a biodiversity specialist at Manchester Metropolitan University. Both researchers emphasize that the infiltration of AI-generated imagery threatens to compromise decades of ecological research and conservation planning. The scale of the threat is underscored by controlled experiments. Lees and his colleagues demonstrated that Google’s Gemini model could easily fabricate convincing bird images or alter existing photographs to mimic distinct species, potentially deceiving both human experts and automated classification systems. Such manipulations are already occurring in practice, with edited submissions previously leading to false species records in regions like Brazil before being flagged by community moderators. Historically, biological databases have managed deliberate hoaxes, but experts note that generative AI acts as a force multiplier, drastically lowering the barrier to creating plausible but nonexistent organisms. The influx of synthetic data strains an already pressured scientific landscape. Taxonomy faces a chronic shortage of specialists, while citizen science databases like iNaturalist and eBird handle hundreds of millions of observations critical for tracking climate change and biodiversity loss. To counter this, platform operators and researchers are advocating for a multi-tiered defense strategy. iNaturalist has integrated community reporting tools to flag potentially synthetic media, while experts like Lees call for standardized image-authentication protocols and rigorous metadata verification. Public education on the risks of AI manipulation remains a foundational component of the proposed response. Despite the security challenges, industry stakeholders emphasize that AI itself is not inherently detrimental to ecological science. Discriminative machine learning models continue to enhance conservation efforts, successfully predicting migration patterns, monitoring algal blooms through the CitClops initiative, and processing camera trap data for the Australian Wildlife Conservancy. These applications rely on AI to sort verified human-submitted data rather than generate it. Maintaining this distinction is critical. As synthetic media grows more sophisticated, the integrity of global biodiversity databases will depend on robust technical safeguards, transparent verification workflows, and the sustained oversight of dedicated citizen scientists who remain the primary bulwark against data contamination.