School deepfake ransoms

Episode Summary

Executive Summary: This episode examines how AI deepfakes are being weaponized against children and schools, including extortion schemes, and explores a possible mitigation: removing real children’s identities from images before publication. It then shifts to a graphene-based bioelectronic chip that borrows animal scent receptors for advanced sensing, and closes with a student’s AI-powered accessibility device for dyslexic users.

Main Topics: AI deepfakes and child sexual abuse imagery (Priority: 5/5): The programme opens with a disturbing report on AI-generated explicit images of children, including cases created by classmates or criminals using scraped social media and school website photos. School extortion and online safeguarding (Priority: 5/5): Criminals are reportedly using deepfaked school images to blackmail institutions for ransom, prompting warnings about public sharing of children’s photos and the need for stronger safeguarding. A platform-level mitigation: synthetic identities for children (Priority: 4/5): ADOS proposes an AI system that replaces real children’s identities in published images with synthetic ones, aiming to reduce the risk of real images being turned into abuse material. Bioelectronic sensing with graphene and animal receptors (Priority: 4/5): Researchers describe a graphene microchip paired with biologically inspired receptors to detect odors and other signals with sensitivity beyond human capabilities. Potential applications and ethical implications of advanced sensing (Priority: 3/5): The discussion covers uses in healthcare, security, agriculture, and robotics, while noting privacy concerns because improved sensing could reveal personal or sensitive information. Lexo: an AI-assisted tool for dyslexic students (Priority: 4/5): Student designer Angelus Collins presents a plug-in device that combines reading, summarizing, dictation, and grammar help into one workflow for dyslexic users.

Key Arguments: Deepfake abuse of children is growing rapidly and has become a global safeguarding crisis requiring coordinated action from parents, schools, tech firms, and law enforcement. Schools are particularly vulnerable because they often post children’s photos publicly, which criminals can scrape and turn into extortion material. Parents and schools may not fully understand the new risks when consenting to online publication of children’s images. Removing real children’s identities from images before posting can reduce harm, even if it cannot stop criminals from scraping the content itself. Graphene combined with biological receptors can create highly sensitive sensors that emulate animal capabilities such as smell. AI is valuable for interpreting sensor data, but it depends on high-quality real-world data generated by hardware like these sensors. Accessibility tools are more useful when integrated into one seamless workflow rather than spread across multiple apps and interfaces.

Data Points: Deepfake reports from children (first six months of this year): 420 - NSPCC / Internet Watch Foundation reporting on images children believed were faked or manipulated to appear explicit Previous total for 2025 deepfake reports: 397 - The number already exceeded the full-year 2025 total noted in the transcript Deepfaked girls targeted from a school website: 30 - LBC report cited by the programme about extortion using school website images Bitcoin ransom demand: more than US$330,000 - Reported extortion demand for threatening to release child sexual abuse material Deepfake nudes created by two teenage boys: more than 300 - Case described as involving mostly underage girls in a small town and elite private school Child-protection scale: global pandemic / public health crisis - Simon Bailey’s characterization of generative child sexual abuse material Research use case example: proof-of-concept chip - Graphene bioelectronic microchip paired with animal-inspired receptors Lexo creator age: 23 - Angelus Collins, product design student at Nottingham Trent University

Pivotal Quotes: "We are seeing a significant increase in the number of images of generative child sexual abuse material. It is now a global pandemic and it's a public health crisis." — Simon Bailey: He describes the scale and urgency of AI-generated abuse imagery and extortion "We cannot stop criminals from scraping these images... but we can remove real children's identities from online platforms." — Carol Osborne: She explains ADOS’s approach to reducing harm by synthetic replacement of children’s identities "We're sort of borrowing the supersensors that nature has developed over millions of years and using them with the chips to expand our capabilities to detect various things." — Dr. Kiana Aron: She summarizes the idea behind bio-inspired graphene sensing

Implications: The episode highlights urgent demand for stronger child-safety practices online, especially around school images, while also showing how AI and bio-inspired hardware may improve healthcare, security, and accessibility if deployed responsibly.