A roadmap for the industry to mitigate bias in advertising
Diversity has been established as a firm part of the narrative in the marketing industry in Canada as a force for good and a force for growth. In the creative process, diversity, equity and inclusion (DEI) impact both the input – diversity of talent, ideas, experiences and regional representation, and the output – the content, imagery, visuals, and voices. Research has shown that the way the creative reaches consumers often has its own built-in biases.
Nearly $1 trillion was spent on digital advertising globally in 2021, much of which flows through programmatic engines that segment and target specific audiences, sometimes missing large consumer groups in the process.
So, when we create and disseminate content, we need to stop and ask ourselves: who are we excluding?
Over the past year, the CMA Creativity Council has explored many aspects of the impact of DEI on creativity. The existence of bias in advertising is not a new conversation. The question is, how we, as an industry, are looking to identify and mitigate those biases. How do we become better? The great news is that a lot of thinking and work is being done in this area by many brands and agencies to drive education, awareness, collaboration, and action.
In 2021, an IBM Watson Advertising research project confirmed that bias can exist in advertising technology, and showed that mitigating that bias is possible, using artificial intelligence (AI) tools and resources in marketing processes.
At the Cannes Lions International Festival of Creativity 2022 earlier this year, IBM released an open-source toolkit to help identify and mitigate bias in advertising technology. The toolkit, AI Fairness 360, is an open-source solution deploying 75 fairness metrics and 13 state-of-the-art algorithms to help identify and mitigate biases in discrete data sets. A playbook and sample code provide guidance on how the metrics and algorithms in the advertising toolkit can be put into practice.
Resources such as these are laying out a roadmap for the industry to systematically address bias. Organizations utilizing the toolkit may gain a better understanding of the presence and impact of bias on their advertising campaigns, as well as the makeup of their audiences. This is not only a major step in ushering a new era of fairness. It can also benefit campaign performance.
While more industry participation and data collection are needed to better understand the potential impact of bias in marketing campaigns, innovations such as this bring energy and a sense of optimism. Use of and collaboration on the open-source tools, data and AI models can further strengthen open technologies that can help automate bias mitigation.
As we reinvent the advertising technology infrastructure with consumer trust at the centre, we must address bias in the technology that powers the entire marketing industry. It is a business imperative, and it is the right thing to do.
We should all be part of the solution. When the industry demands more, real change can happen.