AI-generated nature images can appear convincing, though they often contain anatomical errors or depict non-existent species. «The Beetle Noticed» relies exclusively on real photographs and verified information, helping readers better understand insects without unnecessary fear or misconception.

Telegram will serve as the primary platform for the project. Additionally, materials will be posted on VK, and short videos will be used to attract a new audience.The media content is distributed across several sections. «Bug of the Week» introduces readers to a specific species, while «Who Is It?» is built using photos from subscribers. The «Myth or Fact» section addresses common misconceptions. Individual posts are dedicated to the ecological role of insects and situations in which they can actually pose a danger.

A significant portion of the «BUG Noticed» material will be based on user observations. Subscribers can submit photos of unidentified insects and ask questions about their behavior.Each submission requires processing. We must examine the photograph, determine the insect’s potential classification, and verify the found information. If the observations become numerous, it is difficult for the editorial team to complete all the preliminary work manually.

An additional AI function is involved in preparing publications. After verifying the information, the system drafts the text, which the editor refines based on the nature of the media.

AI can also analyze audience questions. Recurring inquiries help to understand which insects generate the most interest and which topics require further explanation. However, automation must not replace expert review. Since the exact species cannot always be determined from a photograph, the neural network’s suggestions are considered only as a starting point for research.

The primary source of data for the «Who Is It?» section is photos submitted by subscribers via the Telegram bot.Along with the picture, the user provides the region and the approximate time of observation. Optionally, they can describe where they found the insect and what it was doing. This information helps narrow down the possible identifications. For instance, some species are only found in specific natural environments, and the activity of others depends on the season. After receiving the photo, the system assesses its quality. If the insect is photographed too far away or the necessary features are indistinguishable, the AI prompts the user to take another shot. However, the user is not required to provide exact coordinates; an approximate location is usually sufficient for editorial content.

Entomological reference books and scientific publications are used to verify the information. iNaturalist can serve as an additional resource, which hosts real-world observations of living organisms.AI assists in locating materials related to the suspected species and extracting the information necessary for publication. The acquired data is then reviewed by an editor. Special attention is paid to the subtle differences between similar insects, as superficial resemblance does not always indicate membership in the same species. If scientific sources contain conflicting information, the editorial team must not obscure that uncertainty. Publications must only use verified data or clearly state any existing limitations.

Verified observations gradually form a personal media archive. A card is created for each photograph, storing information about the author and the established taxonomy. The date of observation and region are also included, provided the user entered this information. AI helps organize the archive and locate images that can be used for comparison. For instance, when preparing material on similar species, the system can suggest photos from previously verified observations.

Subscribers' comments and questions help the editorial team determine future directions for their work. If readers repeatedly ask about the same insect, the AI can aggregate these inquiries and suggest a topic for a dedicated analysis. Similarly, survey results are analyzed to show which sections interest the audience and which explanations remain unclear. When processing inquiries, users' personal information is excluded from analytical materials.

Unlike media, where AI is primarily used for generating publications, the «BUG Noticed» project begins its automation process by analyzing real-world observations.