Artificial intelligence (AI) chatbots are leading children into “terrible harm” and can push people towards sexualised and pro-eating disorder content, an Oireachtas committee has heard.Noeline Blackwell, of the Children’s Rights Alliance (CRA), on Tuesday said chatbots and erotic companion tools, such as AI girlfriends, are matters causing rising concern.Blackwell told the Oireachtas Committee on Artifical Intelligence the alliance is encountering “more and more” instances where “chatbots are leading children into terrible harm, including suicide, as has happened recently in the UK”.She said fake, AI-generated material poses “greater risks for children than adults” as they have more difficulty identifying false information.Blackwell was responding to a question from Social Democrats TD Sinéad Gibney about emerging threats, beyond the impact of images generated by Grok, the AI tool on social media platform X.The committee previously heard 244 referrals had been made to the Garda concerning intimate image abuse created by nudification or AI generation software. About half of those related to child abuse imagery, Garda Assistant Commissioner Angela Willis told the committee in March.Olga Cronin, senior policy officer with the Irish Council for Civil Liberties (ICCL), told the committee the “Grok scandal” demonstrated that AI systems can “facilitate the creation and dissemination of illegal and deeply harmful content” at speed and scale.The State and the European Union possess the legislative tools and powers to respond, she noted, adding that “visible and timely enforcement” was needed to ensure public confidence.Cronin said the ICCL had no issue with the banning of “nudification” apps, but the “real problem” was that “most general purpose AI models can create this kind of material”.Blackwell called for the continuous review of the risk level of image generation and deepfake applications, with the EU AI Act having designated this software as of ‘limited risk’ when the technology was less advanced.Her colleague, Alex Murphy, said the CRA believes chatbots should “inherently by deemed as high risk” with some exceptions for the software being designated as limited risk.Murphy said chatbots have been observed pushing people towards suicide, but also “inappropriate and sexualised content, eating disorder content”. She said people develop “psychological dependencies” on these chatbots.Blackwell, formerly the chief executive of the Dublin Rape Crisis Centre, said there is an “awful lot of information being collected about children through school programmes” being enabled by AI.She said a “wide variety” of educational technology is being used in schools. While those programmes may be beneficial, the CRA “don’t see” how children’s data rights are being “properly observed” in the absence of proper licensing systems.“Currently, no system exists for the certification, assessment or licensing of EdTech systems in schools,” she said.“As a result, there is no guide for educators nor is it possible to assess the extent to which the various current programmes comply with children’s rights in this regard.”