WIREDGOV

Ofcom has published its Strategic Approach to AI for 2025-26. The report outlines Ofcom’s plans both for regulating the technology’s deployment across sectors, and for its own use of AI within its operations.

Scanning the potential uses of AI within their regulated sectors, which includes content moderation, telecom network security, and optimising spectrum allocation, Ofcom has set out their key plans related to AI for the upcoming year. In terms of their external work, this includes proposals to:

• Develop additional media literacy resources, including work focused on equipping people with the skills needed to enable them to navigate an increasingly AI-mediated space.
• Work with the Digital Regulation Cooperation Forum (DRCF) to develop a successor model to the AI and Digital Hub.
• Invest in building their AI capabilities through upskilling talent and exploring further how they can leverage AI across their operations.
• Work with the Alan Turing Institute to build a taxonomy for safety technology, in order to provide clarity on the standards industry needs to enable innovation.
• Publish research exploring interventions that could help to respond to harmful deepfakes, such as watermarks, content provenance, and AI labels.
• Explore the need for measures to address AI-based harms. This includes forthcoming proposals on the use of hash matching.

Ofcom have also re-evaluated the key AI risks and established that those outlined in their 2024 AI Strategic Update remain the most relevant to their work. These are: synthetic media, personalisation and security and resilience.

Turning to their internal usage of AI, the report confirms Ofcom employs over 100 technology experts, 60 of whom focus on AI. They are currently conducting over a dozen trials on their use of AI in order to increase productivity, improve their processes and generate efficiencies. Whilst these trials span a wide range of uses across the organisation, the report outlines four examples:

1. Using an AI transcriber to streamline the transcription of broadcast content.
2. Developing a customised text summarisation tool to find patterns within consultation responses, and to match the theme with the corresponding documents.
3. Improving spectrum planning, with the potential to improve spectral efficiency and usage in sharing licences.
4. Using AI tools during the software development process.

Within the report, Ofcom looks to how they can support AI innovation in their regulated sectors. This is divided between four areas of focus:

1. Investing in hands-on and technical AI research, including through sandboxes and technical labs. Associated projects include the DSIT-funded ‘spectrum sandboxes’, Online Safety Technology Lab, SONIC Labs in collaboration with Digital Catapult, and research on the application of machine learning for more efficient and reliable propagation models.
2. Providing large data sets to help train and develop AI models. Ofcom has committed to publishing all the non-confidential research data they collect, and has invested in generating unique, large spectrum data sets which they publish as Open Data.
3. Collaborating with other institutions to provide regulatory alignment, including the DRCF, Alan Turing Institute, and the National Cyber Security Centre, as well as a range of international stakeholders.
4. Building relationships and providing active support to stakeholders to increase regulatory clarity. Ofcom will i) support the provision of further advice and support for innovators, ii) develop their understanding as to how big platforms use AI to improve safety on their services, iii) work with the Government on how to build on the success of the UK media industry, and iv) focus on the opportunities presented by the recent growth of AI assurance services providers.

techUK welcomes this refreshed strategy from Ofcom and looks forward to engaging with them on their forthcoming work on AI.