IBM’s AI fix for ad bias promises an inclusive future


An AI solution to advertising bias? Is it possible? On June 20, that’s exactly what IBM promised, and big business promised to unite to improve the equity of marketing campaigns. Companies like WPP, Delta Airlines, the 4As, Kellogg and Mindshare have decided to adopt IBM Advertising Toolkit in an effort to eliminate the biases of digital marketing.

The industry giants gathered at the 2022 Cannes Lions International Festival of Creativity included agencies, brands and other leaders who are committed to fighting bias in advertising. IBM research has shown that it is possible to reduce bias using AI tools and resources. In 2021, IBM had launched a study to understand how bias can seep into campaigns and its effect on consumer trust. They used AI to understand how deep-rooted the problem is and how an AI solution to advertising bias will help digital marketing campaigns. At the time, the company admitted that more effort was needed to develop tools that will effectively eliminate inherent biases and raise awareness of these issues.

Studies have repeatedly shown that biases have been built into the code and also into the datasets used to train AI. Built-in biases are dangerous because they produce results that reinforce the bias and create a confusing loop of expected feedback. Safiya Noble’s book Oppression algorithms highlights how Google’s algorithms, the rules that govern the world’s most popular search engine, are far from neutral and actually reinforce societal racism and sexism. The technology itself is not to blame as it is man-made. On the contrary, it is extremely important to realize the central problem and to develop solutions that reduce the prejudices at the root.

The IBM Advertising Toolkit is an extension of the company’s advertising equity commitment that promises to collect facts and explore the impact of bias, and promote advertising equity. In a written statement, Bob Lord, IBM’s senior vice president for The Weather Company and Alliances, said, “While the risk of bias in advertising is well known, by making this pledge, these organizations are among the first of the sector to act. He applauded their efforts, adding, “Together, we agree to educate ourselves and our companies and ask other industry leaders to join us in helping to reduce bias in advertising.”

Demonstrating their seriousness in dealing with this deep-rooted problem, IBM’s toolkit for combating bias called Advertising Toolkit for AI Fairness 360 has been made available free of charge. According to the New York-based company, it’s “an open-source solution deploying 75 fairness metrics and 13 state-of-the-art algorithms to help identify and mitigate bias in discrete datasets.” Organizations that deploy the toolkit will be able to better understand bias in ads and the composition of their audience.

Mark Read, CEO of WPP, one of the first companies to adopt IBM’s AI solution to combat advertising bias, said that if “used correctly, data can help brands personalize the consumer engagement and identify the most relevant touchpoints. However, we know that biases can exist in algorithms or technology, and that’s why we help our clients assess how and when to use data in a meaningful way that will benefit the customer experience.

According Salesforce 2022 Connected Client Statusr survey, nearly 62% of consumers surveyed said they were concerned about AI bias. In 2021, nearly $1 trillion was spent on digital advertising globally and that number will only increase from now on. As digital marketing targets specific audiences around the world, marketers need to be careful to stay relevant and avoid offending potential customers. An inclusive campaign will help businesses reach more audiences while validating their choices and helping consumers connect with relevant products.


Comments are closed.