Who: The Advertising Standards Authority (ASA)
Where: United Kingdom
When: 17 November 2022
Law stated as at: 12 December 2022
What happened:
Technology is transforming how the ASA regulates misleading, harmful and irresponsible ads. Adam Davidson, head of data science, has announced that the ASA is increasingly using artificial intelligence (AI) to monitor ads, and improvements in machine learning could help to further revolutionise ad regulation.
Digital advertising is happening on a huge scale, on a variety of different platforms like social media, apps, websites, as well as the traditional TV platform. In addition to this, a lot of content nowadays is targeted, based on location, age or interests. For the ASA, this means that they are expected to monitor a huge number of adverts on differing platforms with unique functionalities. Therefore, without the assistance of machine learning and AI, regulating ads properly would not be possible.
Computers are becoming more and more intelligent in that they can interpret images and text, as well as carrying out complex tasks. This technology will enable the ASA to deliver regulation at greater scale and at faster speed.
The ASA data science team are working closely with the compliance team to create tools that capture ads from different sources and then use machine learning to search for misleading text or identify problematic adverts. By using new technology, computers can deal with the majority of adverts and automate a previous manual and labour-intensive process. The most complicated and relevant ads would then be escalated to a human to assess whether they breach the rules. As a result, the ASA can better respond to concerns and continue to provide a high-quality and quick service.
The ASA is already using these types of tools in influencer marketing and cryptocurrency ads. The ASA Annual Report 2021 demonstrates how AI is being used to tackle unlabelled influencer advertising. AI monitors and assesses social media stories produced by high-risk influencers, who frequently do not label their content as ads. This technology enables the ASA to speed up the identification process and take relevant action against those in breach of advertising rules. Furthermore, the ASA is monitoring a large swathe of cryptocurrency ads on a daily basis. Machine learning and AI helps to streamline this process, as well as sharing potentially non-compliant ads with compliance experts.
Looking to the future, the ASA is considering how automation can help the complaints process. Assessing complaints requires detailed understanding of interpretation and context – which are qualities for which AI is not known. However, the ASA strongly believes there is a role for AI here to assist experts in making quick decisions. While the amount of complaints has increased rapidly as digital advertising expands, the ASA is not willing to compromise on its core principles of fairness, transparency and consistency, so we expect to wait a little longer for transformation in this area.
Why this matters
The ASA is using technology in an innovative and world-leading way in order to tackle misleading and irresponsible ads. Its continuing investment in this area and strengthening of its data science team highlights its commitment to transformation.
The advertising regulator’s investment in machine learning and AI demonstrates its priorities and key objectives over the next few years and shows businesses just how serious it is about ad regulation in the digital era.
