Twitter has called for a “collective effort” to tackle brand safety online, as it recognises that more collaboration between platforms is needed to efficiently combat the spread of harmful content.
At Mediacom’s Blink conference on Friday (31 May), head of agency development for Southeast-Asia Chris Bell's message to marketers was that brand safety should be “priority number one”.
“Viewability and ad impact are super important, but brand safety offers the biggest risk to advertisers,” he said. “For Twitter it comes down to ensuring we have the right technical solutions, such as AI and machine learning—plus human review—at scale globally, while allowing advertisers to create partnerships with large media companies.”
While Twitter has made great strides in ensuring brands only appear next to brand-safe content, Bell admitted that “further work” was needed to combat the spread of harmful content online, which is why the company has signed the “Christchurch Call”.
Twitter chief executive Jack Dorsey joined representatives from Facebook, Amazon, Google, and Microsoft at a meeting in Paris last month with New Zealand prime minister Jacinda Ardern and French president Emmanuel Macron to pledge to combat the spread of extremism. Ardern announced the project in the wake of the Christchurch mosque attacks on 15 March.
Bell said greater collaboration like the “Christchurch Call” will help shift the dial on the broader brand safety challenge.
“From an industry point of view there needs to be more of a collective effort around making [brand safety] a priority and ensuring there are the right systems and learnings in place," he said. "Each platform is going about this slightly differently with differing success."
When queried about whether Twitter could introduce a way to pre-emptively flag non-brand safe content before it gets posted, Bell said the platform has to strike a careful balance with freedom of speech.
“We have algorithms and checks in place that take down content that isn’t brand safe," Bell said. "But the nature of platform is real-time and open—there is a big area between extreme content down to what you perceive as being brand safe versus what someone else doesn’t.
“We are making a lot of strides forward whilst keeping nature of the platform intact, which is designed to be an open conversation platform.”
Twitter uses a combination of machine learning, human review and placement control to ensure ads are being shown in brand-safe environments across the platform.
Visual and text-based machine learning algorithms review content and contextual signals to recognise potentially sensitive content. A specialised team reviews and evaluates content to further train the AI model. The third layer allows advertisers to place controls on where campaigns can run.
“Artificial intelligence and machine learning cannot completely ensure a brand-safe environment at this stage, so we combine it with human review,” Bell said.
Twitter has also invested in growing its premium partnerships with media companies to allow advertisers to serve ads next to TV-quality content using its In-Stream Video Solution.
Twitter has verified 950 publishers across the world as brand-safe, and has around 60 video-content partnerships in the APAC region.
Twitter was found to be the most brand-safe platform for advertisers in 2018, according to a survey by GumGum of 274 executives at brands, agencies, publishers and tech providers.
Around half (45%) of respondents voted for Twitter as the top social media platform in terms of brand safety, followed by Facebook at 25%, Instagram at 17%, and YouTube and Snapchat at just 0.38%.