The University of Bath is partnering on a development project with the British Board of Film Classification (BBFC) to explore the power of AI solutions in content classification.

Driven by the BBFC’s vast dataset and unparalleled human insight from over a century of viewing content, and harnessing the power of Amazon Web Services’ (AWS) cloud services that support AI technologies, the project will complement the BBFC’s existing classification and compliance processes.

Building on the BBFC’s significant and long-standing classification expertise, the AI solutions are designed to bring efficiencies to the end-to-end human viewing process.

Partly funded by Innovate UK, the BBFC is developing a prototype for a bespoke AI tool that will identify and tag content issues, and is working in close collaboration with the AWS Machine Learning Solutions Lab and AWS Professional Services teams. AWS have provided their machine learning (ML) and natural language processing (NLP) technologies to develop this multimodal tool.

Experts from the University of Bath’s Department of Computer Science will be involved to maintain the highest levels of data ethics and to mitigate data and algorithmic bias.

However BBFC’s recently expanded team of compliance officers will continue to be at the centre of classification for years to come. Their expertise will always remain essential in ensuring the BBFC’s classification practices remain in step with the expectations of UK audiences.

Dr Julian Padget, from the Department of Computer Science at the University of Bath said: “As a Department, we are delighted to be supporting the team on this exciting project. What is particularly timely about this is the opportunity it provides to put into practice the principles set out in the emerging standards at International Standards Organisation and Institute of Electrical and Electronics Engineers. The formation of the BBFC’s AI Ethics Committee is a clear step in the BBFC’s ongoing commitment to establish policies and procedures concerning AI ethics, and we are pleased to contribute to those activities. We look forward to continuing to share our expertise to ensure that the project’s processes and outputs are benefiting from the latest thinking on the consideration of bias in data-driven systems.”

The alpha testing phase is currently underway and demonstrates promising initial figures, with the tool achieving 80% accuracy when identifying content issues. Initial statistics also highlight the potential for this tool to bring large-scale efficiencies to compliance teams by delivering possible time-savings of 60%.

Additionally, the BBFC is building a separate tool to determine and assign international age ratings, which can be used in conjunction with the tagging tool. This scalable solution aims to enable distributors and streaming services to obtain age ratings and content advice for multiple territories in a streamlined and cost-effective way. Both of these tools share the ultimate goal of driving down the cost of classification in the future.

The BBFC’s classification standards are underpinned by published Classification Guidelines, which set the foundation for all classification decisions and are informed through research with more than 10,000 people across the UK to ensure classification decisions reflect current UK societal standards. The guidelines are updated every 4-5 years and the BBFC is consulting on its guidelines this year, with any changes required by the research coming into effect in early 2024.

Natasha Kaplinksy OBE, President of the BBFC said: “This announcement marks an incredibly exciting transition for the BBFC, as we continue to embrace new technology to help us achieve our core mission of supporting people and families across the UK to make informed viewing decisions about the content they consume. We recognise that in the coming years, Artificial Intelligence (AI) is likely to play a part in a blended approach to classification. Through this project, we are exploring ways in which AI might support and enhance the BBFC’s classification work, including online. I’m proud to be leading such a cultural institution of over 110 years into the future of classification.”