Google’s Jigsaw Announces Toxicity-Reducing API, Perspective, is Processing 500M Requests Daily

Jigsaw, a unit within Google that forecasts and confronts emerging digital threats, announced that Perspective, a free, open source API that uses machine learning to spot toxic comments, is now processing 500 million requests daily and is measurably reducing toxicity to make conversations online better at scale. The third issue of the company’s online research magazine, The Current, which was published today, details the persistent and damaging issue of online toxicity and the role machine learning tools, like Perspective, can have in reducing it.

Read More : Alida Welcomes New Chief Revenue Officer to Accelerate Global Growth

Launched in 2017 to combat online toxicity and harassment, Perspective scores comments based on their perceived impact on conversation and helps platforms keep comments sections open for engaging and productive conversations.

“Toxicity on the internet is a pervasive problem that disproportionately impacts marginalized groups, threatens independent journalism and crowds out freedom of expression and healthy dialogue,” said Jared Cohen, Jigsaw’s CEO. “Our partners have seen Perspective measurably reduce toxicity on their platforms, enabling them to create safer spaces for their users to exchange ideas. We’re committed to working with our partners and academic institutions to continuously train and retrain our models to become even better at identifying toxicity while minimizing bias in support of healthier conversations.”

Read More : SalesTechStar Interview with Sam Zayed, Chief Revenue Officer at Conga

Jigsaw built Perspective using models developed in partnership with Google’s Counter Abuse Technology team, and it is now empowering a variety of platforms to create safe space for dialogue through multiple uses:

Commenter Feedback: Perspective can provide feedback to commenters about the potential perceived toxicity of their content in real time, leading to significant results:

1. Coral by Vox Media, a prominent commenting platform used by media organizations around the world, uses this function to give commenters an opportunity to change their wording before they post. A controlled study with McClatchy found that 40 percent of users changed their comment in response to feedback.

2. OpenWeb, a leading audience engagement platform that hosts 100 million active users per month, launched this feature after an OpenWeb study, conducted as part of broader Jigsaw research, found that out of all commenters who edited their language, 44.7 percent of commenters replaced or removed offensive words and 7.6 percent elected to rewrite their comment entirely.

Read More : Sales Recovery: Road Map to Post-Pandemic World