What’s in a word: Challenging keyword biases to make brands more inclusive

What’s in a word? The female protagonist of Shakespeare’s tragic love story might not have believed that names, and words, matter but today, with inclusivity top of the marketing agenda, words have meaning.

The Drum and Shutterstock roundtable brought together leading marketers to discuss how brands can help challenge biases by changing their metadata and searchable keywords to break down stereotypes and create a new, more inclusive world.

All stereotypes start with a search. From Microsoft using AI to suggest inclusive words, to ASOS relaunching ‘beauty’ as ‘face & body’, to Zalando’s ‘Zerotypes’ campaign as an antithesis to stereotypes - brands are starting to replace outdated cultural stereotypes with more inclusive terms but there is still a long way to go in addressing the biases within search. The Drum and Shutterstock roundtable participants agreed that before we reset the algorithms, real change needs to start with the role that humans play.

Search biases are not, and have never been, just about technology

“We have to look at this in every way - who are the humans creating the technology? Can we build an alternative to start working that out? Can AI talk to AI to figure this out as we move forward?”, asked Mordecai, head of innovation + partnerships, Innocean Worldwide Americas (Hyundai Motor Group). After all algorithms do what they’re taught; and unfortunately, some are taught prejudices and unethical biases by societal patterns.

“It’s fundamental to look at who’s creating and developing that space and make sure we have awareness and a diverse group of people having those conversations. The issues really come when the human input is only one sort of human.”

In building this machine learning system, what then is an inclusive and comprehensive search approach?

Technology has come a long way, yet search is “not super smart yet”, according to Kristen Sanger, senior director, contributor marketing, Shutterstock, noting that computers still struggle with contextual search.

According to Sanger, “How we describe people, things and colors can start to intertwine within search, which struggles with contextualizing what the keywords are saying within the content. Human intervention ensures that metadata is appropriate and correct for what humans are going to search for, not things that computers might automatically detect.”

Humans vs machine

Human biases need to be tackled before we even touch on technology to ensure the inequities in search are set right, added Steven Moy, chief executive officer, Barbarian.

“Humans still need to teach AI to learn about contextualization, persona development and understand different countries, tonality and context.” The problem, according to him, is that simple algorithms treat all data as immutable, even data about our preferences, income, life situation or countless other patterns that continue to shift. Now is the time to move on from historical indexing engines based on outdated keyword and metadata and “remember that the explosion of digital transformation in AI and machine learning can really teach [search] to know the differences.”

So, if algorithms can trap people in their origins, or a stereotype is there a way to combat it all?

“We need to think about debiasing humans as much as debiasing technology and systems,” said Dipti Bramhandkar, executive planning director, North America, Iris Worldwide. “Working on both fronts will enable us to create a fairer understanding of how search works but also how we relate to each other and the world that we live in.”

Recreating a new narrative

Paola Ortega, associate strategy director at DDB, noted three things to build: awareness, responsibility and a new narrative: “Start to build a new narrative by breaking down the stereotypes and building empathy with consumers. All of that then trickles down to how creative is influencing society and how we fight for more inclusion.”

It begins with asking ourselves - how do we question our own prejudices, not just in the tech we build, but in the way we treat those around us?, the panellists agreed.

Cassie Begalle, strategy and innovation brand manager- U by Kotex, added that brands have a huge role in ensuring diverse opinions are driving these conversations: “We have to unstrip and understand the biases in our brands, and also have the courage to start breaking category biases.

“Ideally, we want to get to a point where agencies and brands are okay challenging each other and working through these issues together so we can make sure that we’re coming at this the right way,” added Begalle. “We have to make sure that the fear of failure doesn’t hold us back from making progress. We are the champions for consumers and we need to make that change.”

Destroy to create

Moy urged brands to be courageous; be experimental; and “destroy to create” He explained: “You have to unlearn everything you do in order to create a new way of thinking.” Moy said there need to be systems in place for holding ourselves and each other accountable for our behaviour, which would be a solid first step to ensuring that biases are mitigated and ultimately, reduced.

It comes back to not just the fundamentals of marketing but of humanity and authenticity. “Ask a brand if they would ever like their blocklist to be shown publicly?” urged Mordecai. “What are we doing to be proud of the values we have and express them? Blocklists may not be on everyone’s cancel culture list right now, but right now we have this opportunity.”

Begalle added that both agencies and brands have the responsibility to address our own and each other’s prejudices and ethics.

“As brands we need to take a step back and audit our exclusion lists and negative keywords. And make sure that we’re doing right by our consumers because sometimes it is very tactical, and we don’t examine it as much as we should. But it is the foundation of how we can start breaking these stereotypes. Sometimes we need to shift our priorities to make sure important things affecting an impact change are going to happen.”

And real change can only happen when the people guiding the discussions have typically been “the minority in majority cultures”, concluded Bramhandkar, versus incremental changes with the same people in the same way using the same systems.