July 29, 2021

AI for the detection and mitigation of advertising bias

IBM Watson Advertising today announced a research initiative that will apply IBM’s open source AI technology to better understand the prevalence of unwanted bias in advertising, while simultaneously laying the foundation for potential mitigation. This work aligns with IBM Watson Advertising’s mission to make AI the catalyst to improve advertising solutions and services. In addition to reliance on the advertising ecosystem, and builds on IBM’s leadership in business transformation across industries with AI.

The advertising industry is undergoing a major transformation, with changes in privacy policies and an increased demand for trust and transparency. All of which fuels the rise of AI as the new technology backbone of the advertising industry. As brands rebuild in the midst of this transformation, it is the ideal time to address the persistent problem of unwanted bias in advertising, as evidenced by a study by the Geena Davis Institute on Gender in Media, which revealed that in male characters appear 12% more than female ads. Meanwhile, gender-balanced and female-led videos produced 30% more views than other videos. Which reveals a demand for more inclusive content.

Unwanted bias in advertising has the potential to negatively impact consumers

Well, they may miss out on a potential financial opportunity or feel targeted due to stereotypes. While it also hurts brands as they may experience poor campaign performance. The goal of the IBM Watson Advertising research is to drive a deeper understanding of the factors that contribute to bias in advertising and how AI can help. Covering audience segmentation, creative messages, performance optimization and campaign impact. Equipped with this information, marketers and IT vendors can develop a roadmap for the responsible use of AI. This is to reduce bias and create and run campaigns more effectively.

“We are at a moment in history in which a discourse on socioeconomic inequalities. That it should have happened a long time ago dominates the national agenda and calls for action ”. Commented Bob Lord, Senior Vice President of Global Ecosystems and Blockchain at IBM. “Our hope is that artificial intelligence can be a catalyst to eliminate unwanted bias in advertising. Just like you are helping transform the industry to prosper in a cookie-free era. Through this research, we are taking an important first step toward that goal. Applying scientific rigor to determine the dimension of the impact that AI can have ”.

IBM Watson Advertising will work with researchers from IBM Research to conduct this study. This in partnership with the Ad Council and other industry and academic leaders.

The exploration areas will include:

  • Incidence of bias in advertising: Prevalence and frequency of biases in campaigns through performance data analysis. For example, using the AI ​​Fairness 360 Toolkit. A set of open source AI tools developed by IBM and donated to the Linux Foundation. The study will analyze how certain audiences from past and active campaigns are being targeted with creative content to assess whether bias was present.
  • The role of signals in determining biases: The intensity with which the signals, which refer to the context in which an advertisement is delivered, impact the bias. For example, if a creative message is considered to be unbiased in itself but delivered on a digital media property associated with inherent bias, the ad may be perceived as biased.
  • AI capabilities to potentially mitigate bias: how useful artificial intelligence can be in identifying instances of bias and what can be done to harness the full power of artificial intelligence to potentially prevent the occurrence of ad bias.

Data from the educational campaign on the COVID-19 vaccine It’s Up to You (“It’s Up to You”) from the Ad Council will be used for this initial investigation phase.

IBM’s extensible open source toolkit. AI Fairness 360 will also be used to examine the data. And also determine if there are ways that artificial intelligence can be useful to help mitigate discrimination and bias.

“Collective bias has been persistent in our industry for quite some time, and the need to study its origins and impact is critical. This is so we can work together effectively to create progress, ”said Lisa Sherman, President and CEO of the Ad Council. That’s why the Ad Council is proud to be the industry’s first partner to work with IBM. This to help drive this fundamental investigative mission ”.

IBM Watson Advertising plans to publish initial research results and updates as they become available and as the initiative progresses.

If you don’t want to miss any details about video games, technology, anime and series, stay tuned for our news section.