Rebalancing toward proactive ad regulation for the digital era

Our Annual Report 2021, published today, highlights the innovative and world-leading ways in which we have harnessed technology, in line with our strategy, to tackle misleading and irresponsible ads online.

Coinciding with the report, we are today also announcing how our use of artificial intelligence (AI) is revolutionising and enhancing our monitoring capabilities to help identify and take enforcement action against influencers who fail to disclose when their posts are ads.

The Report also includes:

  • For the first time, a breakdown of complaints by nation, providing us with a clearer picture of national differences alongside the overall complaint totals.
  • The key themes and issues – from environmental claims, body image, harmful racial and ethnic stereotypes and cryptoasset advertising, to age-restricted ads seen by children – that we’re shining a regulatory spotlight on to ensure the public is protected.
  • Details of our Intermediary and Platform Principles Pilot, a world first involving some of the largest companies in the digital advertising supply chain, that extends the ASA’s role online and aims to improve transparency and widen accountability in online advertising.
  • Facts and figures including the record number of complaints we received last year, the ad campaigns complained about and the media in which they appeared.

Using to AI help tackle unlabelled influencer advertising

We are announcing a dynamic new project that harnesses artificial intelligence to capture and analyse all Instagram Stories produced by ‘high-risk’ influencers who are on our radar for being unwilling or unable to clearly and consistently label when their content is an ad. Using machine learning algorithms to automatically identify potential ads in these posts, we are taking advantage of state-of-the-art image recognition and natural language processing techniques to categorise the images and videos influencers post. We also search influencer’s content for possible ad labels giving us the ability to identify which are likely to be ads but are not disclosed as such, in breach of our rules.

The technology is transforming our ability to act at pace and scale in identifying problems and take action. It automates what was previously a manual and labour-intensive monitoring process, resulting in efficiency gains, and enables us to prioritise where we focus our attention – on influencers and advertisers causing most harm to consumers.

The immediate results from this project are:

  • We are now monitoring dozens of high-risk influencers each week
  • We have captured and analysed more than 80k Instagram Stories for potential non-compliance since the project started
  • We are now capturing and analysing nearly 20k Instagram Stories each month

We are eyeing improvements, scaling up this process, tracking what influencers are talking about and examining trends in ‘tagging’ and ‘mentions’. We are also ramping up our monitoring of companies who partner with influencers and holding them to account. Repeat offenders who are unable or unwilling to follow the rules will face an escalation of sanctions and enforcement action. As part of that, we are currently considering lining up the worst offenders for referral to statutory authorities for consideration of legal sanctions as well as working directly with the host platform to pull the plug on problem accounts.


2021: Key facts and figures

The Annual Report provides a breakdown of the complaints and cases* handled by the ASA as well as the advice, training and compliance work undertaken by CAP over the last 12 months. In 2021:

  • The ASA received a record 43,325 complaints (about 22,115 cases)
  • Of those:
    • Online remained the most complained about advertising media making up close to 1/2 of all complaints (20,735) but almost 2/3 of cases (14,558)
    • Complaints about online ads were up 19% on 2020 while case numbers remained at the same level as the previous year
    • TV was the second most complained about advertising media also making up close to 1/2 of all complaints (20,425) but only 1/5 of cases (4,802)
    • TV complaints increased by 44%** on 2020 but there was a 5% decrease in the number of TV ads complained about
    • Complaints about influencer posts (4,889) increased by 20% and made up almost 1/4 of all cases (3,648), a 9% increase
    • There was an increase in complaints and cases about other online ads*** including:
      • Video on demand (138% increase in complaints and a 22% increase in cases)
      • Paid-for ads on websites, social media and apps (39% increase in complaints and a 14% increase in cases), and
      • Audio podcast or on demand (54% increase in complaints and a 9% increase in cases)
    • There was a decrease in complaints and cases for:
      • Website, social media or app (own site) (1% decrease in complaints and 9% decrease in cases)
      • Game (a 38% decrease in complaints and a 37% decrease in cases)
    • 20,456 ads were amended or withdrawn as a result of the ASA’s work
    • CAP delivered 866,145 pieces of advice and training to businesses on the advertising rules
*A case refers to an ad/ad campaign we have received complaints about.  One case can have many complaints
** A Tesco ad featuring Santa with a Covid passport prompted 5,064 complaints making it the second most complained about ad of all time
*** Online advertising is a broad category and includes, among others: companies’ own websites, social media and apps, influencer ads, paid-for online ads, video on demand, search and game.


Complaints by nation

For the first time we are also reporting on complaints and cases broken down across the UK. We’ve analysed the figures for each home nation as a percentage of the UK population, providing us with an insight into distinct variations between the nations.

Our analysis reveals that:

  • Proportionately, people in Scotland are more inclined to lodge a complaint about an ad, whether on the grounds of misleadingness, harm or offence, than people in any of the other nations
  • As a proportion of the UK population, people in England complain about ads close to their population level
  • People in Northern Ireland and Wales are less likely to lodge complaints about ads
  • While people in Wales were less likely to complain, they were more likely to complain about issues relating to harm and offence in ads than people from England and Northern Ireland

As this is the first year we’ve run these figures there is no comparative data that enables us to identify any mid- to long-term trends. While the reasons for complaints from Scotland being higher at a population level are unclear it could be explained, in part, by a national ad campaign the ASA ran across Scottish Media between September 2020 and March 2021. It resulted in a statistically significant increase (56% v 50%) in Scottish adults who are now more certain that an organisation exists to complain to about advertising driven by those who had seen the ASA’s ad (61% certain).


ASA Chief Executive, Guy Parker says:

“Technology is transforming all our lives and as our Annual Report shows it’s also transforming how the ASA regulates misleading, harmful or irresponsible ads. As our world leading use of artificial intelligence to help tackle misleading influencer ads demonstrates, we’re harnessing and increasing our use of artificial intelligence and machine learning to deliver tech-assisted and proactive regulation. It means we’re better able to respond to concerns and ensures we continue to provide a one-stop-shop for advertising complaints for the public and responsible businesses across the UK.”

More on

  • Keep up to date

    Sign up to our rulings, newsletters and emargoed access for Press. Subscribe now.