How AI Will Enable More Responsible Advertising In 2023
AI would make advertising more responsible. Find out how.
Today’s consumers expect a brand’s values to align with their own values. This expectation is springboarding responsible advertising and, by extension, brand suitability into the forefront of industry conversation. As a result, advertisers and publishers have placed an increased focus on how best to achieve the goal of strategic, responsible advertising. However, with the rise of digital media, managing where and what type of content advertisements appear next to has become increasingly difficult. Aaron Andalman, Chief Science Officer and Co-founder of Cognitiv, discusses how AI will help by enabling brands to focus on more responsible advertising.
An unfortunate side effect of programmatic advertising has also been a plague of fraud and misinformation, leading to ad placements that are adjacent to values or content that may not necessarily resonate with those of the advertiser. These types of placements can be detrimental to a brand’s image prompting customers to believe that the brand supports messages they don’t or unintentionally causing harm to their customers by amplifying negative or misinformation.
Finding ways to avoid these types of ad placement has proved challenging because programmatic advertising requires making millions of placement decisions every second. While it has been possible for brands to avoid entire domains/publishers that consistently fail to align with their values (using coarse tools like exclusion lists), this system has proven inadequate for aligning ad placements with brand values. Fortunately, the year ahead looks promising as emerging technologies introduce opportunities for the industry to make more responsible and intentional ad placement a reality.
See More: Using AI to Enhance Video Marketing Strategy — Customer Experience
What Is Responsible Advertising? Why Is It Challenging?
Truly responsible advertising is built around ensuring your approach to advertising and marketing positively impacts society, the environment and more. Adopting ethical advertising practices that promote a brand’s value and mission while also shifting spend away from negative or harmful content makes this possible. One of the most significant challenges facing the responsible advertising movement has been enabling a more tailored, accurate approach to assessing the quality and meaning of an individual web page.
While a human can easily evaluate whether a webpage aligns with a brand’s values, there are too many web pages for humans to manually read and score all of them. The common approach to address this issue is to combine keyword analysis with traditional machine learning, but these solutions have had limited success. Identifying a responsible placement for an ad requires much more than keywords, block lists and filters. Research shows that more than half (57%) of neutral or positive stories on major news sites are being incorrectly flagged as unsafe for advertising. This is largely thanks to vast exclusion lists blocking anywhere from 2000-3000 keywords that are defined as “unsafe” without considering the context in which they are used. The reverse is also true when evaluating image-based content such as memes, which may contain text that is benign at first glance but becomes offensive when read with the images’ added context.
While today’s solutions have helped ensure brand safety, i.e., preventing brands’ ads from appearing near media that has the potential to damage their brand, the limitations are significant. This includes limiting campaign reach, reducing relevant programmatic ad inventory and income for publishers, and impeding advertisers’ ability to optimize their media strategy effectively.
Improvements in AI Make New Solutions Possible
While previous iterations of machine learning and natural language processing were limited in their ability to answer complex questions about content like “does this content align to my brand’s values,” new advancements have given the technology a more human understanding of context, allowing them to align content to values more effectively.
The technology underlying these advancements are large foundational language models. These models trained are using enormous corpora of text scraped from the internet. These data sets are then used to train equally enormous artificial neural networks that push the limits of what is possible with today’s supercomputers. The result is computer models that can understand and generate text far better than any prior technology. The models are so impressive that some scientists have controversially postulated the models are sentient! While this may be a stretch, these models clearly open entirely new possibilities for the types of tasks computers can accomplish. However, deploying these models on real-world business problems is difficult.
Large language models are computationally demanding, and they often have to be fine-tuned using specific data before they can be applied to a particular use case. Over the past couple of years, forward-thinking businesses have been busy developing the training data sets, infrastructure, and business logic needed to bring this technology to market, which will accelerate in 2023. The increasing availability of this AI technology in 2023 will empower advertisers to address their unique contextual and suitability needs more accurately, more easily avoiding content they do not wish to be associated with for a more responsible approach to advertising.
See More: How Is Machine Learning Transforming Video Advertising
With a more subtle and accurate understanding of the context in which ads are placed, advertisers will be able to more successfully avoid media buys on web pages that are non-inclusive, contain disinformation or fail to align with their core values. As new research and infrastructural investments bring this AI technology to market, new and improved products will meet the advertising industry’s desire to make more thoughtful, strategic and personalized media decisions. In particular, better contextual targeting tools will allow programmatic advertisers to better avoid spending on web pages that do not align with their values.
Using AI to make smarter media buying and ad spend decisions will spark an industry-wide shift from brand safety (avoiding any risky content) toward brand suitability (i.e., how well a webpage fits within a brand’s set risk parameters) that will be invaluable in allowing brands to advertise more responsibly across media channels.
Which brands do you think have become more responsible with their advertising? Tell us on Facebook, Twitter, and LinkedIn.
Image Source: Shutterstock
MORE ON AI
