How IBM Watson Advertising Plans To Use AI To Reduce Bias

IBM Watson Advertising kicked off a six-month research project last week with plans to have an overview of the results in late fall, and the project completed by the end of the year. The research focuses on bias in advertising.

Shortly after, the company plans to share the findings, according to Sheri Bachstein, CEO of The Weather Company, and GM for IBM Watson Advertising.

Bachstein said the project may require the use of other data sets, but initially it will analyze data from IBM and The Weather Company’s own campaigns, and from the Ad Council’s “It’s Up to You” campaign, as well as the COVID-19 Vaccine Education Initiative led by the Ad Council and COVID Collaborative used to educate the American public on COVID-19 vaccines.

Search & Performance Marketing Daily (S&PMD) caught up with Bachstein to talk further about the project announced last week.



S&PMD:  What does IBM hope to learn, and what will the company do with the findings?

Bachstein:  The advertising industry is facing major disruption, with changes to privacy policies and increased demand for trust and transparency.

We are seeing some very encouraging signs that AI can be the transformative technology that advertising needs to predict optimal campaign performance across the open web and let brands and agencies buy against those predictions -- all while being fully privacy-forward.

As the industry rebuilds amidst all this transformation, the timing is ideal to address the persistent problem of bias in advertising. In fact, as an industry we have an obligation to address social inequities and to do so not just on the surface.

This research will help us determine if AI can impact and potentially mitigate bias in advertising. From there, we’d like to see if we can help the media and marketing industry develop advertising practices that contribute to a more inclusive and equitable world.

S&PMD:  How is IBM defining bias for this research?

Bachstein:  For this initiative, we’ll be examining bias in three different areas.

  1. Incidence of bias in advertising – the prevalence and frequency of bias in campaigns through the analysis of performance data. Using a Fairness research algorithm from IBM Research, the study will look at how certain audiences of past and active campaigns are being targeted with creative content to assess whether bias was present.
  2. The role of signals in determining bias – how heavily signals, which refer to the context in which an advertisement is delivered, impact bias. If a creative message is deemed to be unbiased on its own, yet is delivered on a digital media property associated with an inherent bias, the advertisement may be perceived as biased.
  3. Capabilities of AI to potentially mitigate bias – how useful AI can be in identifying instances of bias, and what can be done to fully capture the power of AI to potentially prevent occurrences of bias in advertisements.
Next story loading loading..