Member

436: After Clause 213, insert the following new Clause— “Defences to the AI chatbots offencesIt is a defence for any person charged with an offence under sections (AI chatbots: content promoting terrorist and national security offences) or (AI chatbots: offence) to prove that they—(a) created, supplied or otherwise made available an AI chatbot which breached the provisions in those sections for the purposes of the prevention, detection or investigation of crime, or for the purposes of criminal proceedings, in any part of the world, (b) were a member of, employed by, or engaged by OFCOM and created, supplied or otherwise made available an AI chatbot which breached the provisions in those sections in the exercise of OFCOM’s online safety functions under the Online Safety Act 2023,(c) were red-teaming or testing the product, for the purposes of ensuring its compatibility with applicable legislation or guidance, or(d) both—(i) took reasonable steps to avoid committing the offence, including conducting a written risk assessment on the likelihood of the AI chatbot in question producing content specified in those sections, and(ii) provided accessible and transparent means with which users could report content produced by the AI chatbot in question of the kind specified in those sections.”Member’s explanatory statement This amendment establishes reasonable defences for the offences under sections (AI chatbots: content promoting terrorist and national security offences) or (AI chatbots: offence). It is part of a set of amendments related to AI chatbot offences in Baroness Kidron's name.