Keyword blacklists, created in the name of brand safety, have been found to reflect dangerous biases against LGBTQ people, according to a new study from CHEQ.

It is estimated that there are 11 million LGBTQ adults in the US alone. The LGBTQ community has seen progress from brands who have spent large amounts of money targeting this vital community. During Pride Month in June, social media feeds of the world’s most famous corporations switched their familiar logos of red, blue or orange to the vibrant rainbow flag.

However there remains one aspect of LGBTQ marketing that remains strictly black and white. This is the crude business of keyword blacklists, created in the name of “brand safety.” These blacklists are words deemed too dangerous for advertisers to appear beside. If an online news story contains one of these words, then the concerned advertiser steers clear.

In the first study of its kind, CHEQ’s data science team found that nearly three-quarters (73%) of neutral or positive LGBTQ online news stories are being incorrectly flagged by the use of these blacklists.

In the report, CHEQ procured an industry-standard blacklist of 2000 words which are designed to keep brands away from toxic stories. While the list includes words such as “death” “racism”, “sex” and “injury”, blacklisted words also include “lesbians”, “bisexuals", and “drag queens”.

CHEQ ran the full blacklist against 225 neutral or positive online articles on a single news day. This included LGBTQ news sites such as Pink News, Gay Star News and Advocate.

Shockingly, 73% of safe stories on these sites were incorrectly flagged as brand unsafe, effectively de-monetizing this content. This was because of mentions of “cult lesbian thriller, Killing Eve" (keywords: ‘cult’, ‘lesbian’ and ‘killing’), “same-sex romance” (keyword: sex); and US reality TV Show Drag Race ("drag queen"). The Women's World Cup, one of 2019's biggest stories, was also unceremoniously flagged. This was because of a mention that the US and Netherlands’ female soccer teams “featured five lesbian and bisexual athletes." The advertising rug was pulled programmatically from association with the tournament, despite hundreds of brands having clamored to spend $43 million to sponsor and associate with these stars.

Since the study’s release we have been inundated by LGBT executives talking about a daily struggle against blacklists. Benjamin Cohen, CEO of PinkNews Media Group notes: “Saying ‘lesbian’ is a death sentence to advertisers.” Orlando Reece, CEO of Pride Media, speaking to NBC News, says: "Brand safety needs to be a conversation involving people, not technology that is ill-equipped for today’s digital publishing landscape."

LGBTQ campaigner Jerry Daykin, who is EMEA media director at pharmaceutical giant GSK, presented the CHEQ findings to the World Federation of Advertisers. He adds: "We’ll continue to see minority voices squeezed out of media if advertisers don’t actively think about how they spend with them. Brand safety measures often actively exclude LGBT audiences with blocked keywords.”

As an AI-led brand safety platform, CHEQ has a vested interest in the adoption of more sophisticated technology to ensure smarter brand safety that does not exclude such safe content. However, as a basic starting point, we must all agree that brand safety should not be alienating the exact audiences we want to target. (Cohen added that even when brands wanted to explicitly serve ads on PinkNews they were blocked.)

Everyone agrees that genuinely brand unsafe stories (violence, sex, hate, or adult content) should not be monetized. However, along with LGBTQ content creators, such as Pink News and The Advocate, we are campaigning for an advanced solution. CHEQ has enhanced its brand-safety AI to cover 200 granular news categories. The AI can tell the difference between a story about a same-sex lesbian couple getting married, considered brand-safe by most progressive brands, which would otherwise be automatically blocked by most ad verification platforms because it triggers keywords like 'lesbian' and 'sex'. Overall, this advance enables the secure blocking of unsafe content while unlocking greater reach for all campaigns.

This fight has come too late for some. Online news site Gay Star News entered insolvency at the end of July (73% of safe stories on Gay Star News were incorrectly flagged as unsafe in our analysis). Two other LGBTQ sites, Pool in March, and INTO, in January, were also forced to close.

In their parting shot before shuttering, Gay Star News founders Tris Reid-Smith and Scott Nunn talked about the marketing sector having shown "apathy" in its approach. He argued that the LGBTQ community is being denied a vital voice at the exact time when "more people want and need what we offer than ever before.”

Brand safety is not perfect. It takes some effort to ensure contextual understanding of content and the use of new techniques such as AI. But in 2020, surely brand safety can perform much better than this.