Illustration: Lazaro Gamio / Axios
Since the day after George Floyd's death on May 26, the rate of hate speech online in the U.S., as tracked by one digital measurement firm, has been nearly three times higher than typical.
By the numbers: On June 3, at the height of nationwide protests, DoubleVerify, which uses its own technology to scan pages online so advertisers can avoid objectionable content, says instances of hate speech were more than 4.5 times higher than usual — the highest-ever rate it has measured to date.
Details: States with heavy protests experienced the highest levels of hate speech online.
- Minnesota, Washington D.C., Delaware, New York, Connecticut, Vermont, Washington, Oregon, Colorado, and Virginia saw the highest spikes, per DoubleVerify. Each state has experienced at least a 2.2 times increase in its own average rate for hate speech online.
- Real-world events tend to cause spikes in hate speech on the internet. Race-related controversies or violence, hate crimes and political events all tend to elicit more hate speech, DoubleVerify's VP of marketing Heather McKim tells Axios.
- "This year, several primaries and debates saw spikes in hate speech," says McKim. For example, DoubleVerify found that fake news spiked in response to Iowa's botched Democratic caucus earlier this year.
Hate speech was declining prior to the protests, McKim explains. "Also interesting is that following the outbreak of the pandemic until George Floyd’s death, we actually saw a pretty significant downward trend in hate speech overall."
How it works: DoubleVerify says it uses proprietary technology and artificial intelligence to "identify and block ads from appearing alongside online hate speech." It defines hate speech as content featuring "biased or derogatory language or behaviors directed toward an individual or group, based on their race, ethnicity, gender, sexual orientation, religion, or ability."
The big picture: The spikes in hateful content, especially as it related to race and violence, have put pressure on tech companies to take action.
- Last week, Axios reported that the Global Alliance for Responsible Media (GARM), an industry body consisting of the world's biggest advertising companies — including a few Big Tech companies — has agreed to try to define hate speech across the entire industry."
- The move signified a new step by tech and advertising giants towards a collective approach toward hate speech, even though it's mostly a formality for now.
In a note to advertisers, Facebook's VP of global marketing solutions Carolyn Everson said that the industry, via GARM, has settled on four areas in which it will take immediate action: definitions of harmful content like hate speech, measurement, audits and suitability controls.
- The group says they agree on 11 standard definitions of harmful content that had been recently agreed to by GARM's brand safety working group "with immediate focus on Hate Speech + Acts of Aggression" and plan to align on those definitions next month.
- In her note, Everson said she would be providing an update from Facebook about how the tech giant limits ads from appearing next to "hate speech" or "acts of aggression." She says that meetings have already taken place with Facebook's policy team and GARM on the issue.
Yes, but: Tech platforms still maintain their right to more narrowly define and police hate speech individually.
Our thought bubble: The entire internet, from Facebook and Google on down, is groaning under the weight of hateful posts. Pressed by concerned advertisers on one side and outraged users and employees on the other, tech's giants have tried to take countermeasures, but they haven't been able to keep up.