The gala on a recent Thursday evening was typical boom-time San Francisco glitz. Attendees from Google, Yahoo and Facebook traded in jeans and hoodies for cocktail attire and snacked on hanger steak and new peas inside the Beaux Arts City Hall. The convening cause, however, was rather dark: The 30th anniversary of the founding of the National Center for Missing & Exploited Children. The non-profit worked with these tech giants to install PhotoDNA, which scans all images uploaded to their sites for child porn. This nexus of light and dark is where Del Harvey lives, and there she was working in the room in a sleeveless black dress. “I haven’t seen you since the suicide summit,” she greeted one group of guests.
Harvey was the 25th employee at Twitter, where her official title is vice president of trust and safety, but she’s more like Silicon Valley’s chief sanitation officer, dealing with the dirtiest stuff on Twitter: Spam, harassment, child exploitation, threats of rape and murder. As Facebook and Twitter have become the public squares of the digital age, their censors now “have more power over the future of privacy and free expression than any king or president or Supreme Court justice,” writes constitutional scholar Jeffrey Rosen. Twitter famously prizes free expres- sion, but as a business, it needs to ensure its platform doesn’t turn into a toxic-speech zone that scares off users and advertisers. Harvey is the person Twitter trusts to walk that line. With a daily volume of a half-billion tweets, “your one-in-a-million chance of something going horribly wrong happens 500 times a day,” says Harvey. “My job is predicting and designing for catastrophes.”
Not listening to Harvey tends to be a bad idea. Last December, Twitter decided to eliminate users’ ability to “block” people they didn’t like from following and retweeting their accounts, replacing it with a mute option so they simply wouldn’t see the trolls in their feed. Harvey warned it was a terrible idea and would make cyberbullying easier. The blocking feature was pulled anyway, and the ensuing outcry was such that Twitter reversed the decision within 12 hours. It later tacked on the mute button as an option.
Harvey has an unusual background for someone with so much power over public speech. She isn’t a lawyer and won’t say if she graduated from college. Del Harvey is not her real name. She is secretive about her past but allows that she grew up in the South, where she spent a summer as a lifeguard at a state mental institution working with troubled youth. Her education about the dark side of the internet came instead from experience. In 2003, when she was 21, she started volunteering for Perverted Justice, a group that posed as young kids online to engage potential paedophiles in chats. When they “caught” one, they’d post the chats along with the identity of the would-be molester. She eventually became the site’s law enforcement liaison, bundling up evidence for local police, and later reprised her young-girl (and boy) decoy role on the NBC show To Catch a Predator. Her work put people in jail, and she adopted the pseudonym then to conceal her identity from exposed paedophiles. “I do a lot in my life to make myself difficult to locate.”
It informs her work: She advised Twitter to scrub location data from uploaded photos to prevent stalkers from using them to locate people.
Harvey was hired by Twitter in 2008 to deal with a proliferation of spam accounts harassing early users. “Del became an encyclopedia of the weird things people were doing,” says Twitter co- founder Jason Goldman. Though she accidentally shut down the founders’ accounts as “spam” when she ﬁrst arrived, she proved herself by thwarting the pranksters at chatboard 4Chan from derailing a race between Ashton Kutcher and CNN to be the ﬁrst Twitter user with a million followers. When Goldman left in 2010, his farewell advice was to protect Twitter’s brand by protecting users and “respecting their voice.” He wrote, “In case of emergency, trust Del.”
Twitter is so big that Harvey’s decisions invariably offend someone. The site came under criticism this year for blocking tweets in Pakistan—it later unblocked them—and last year in the UK for allowing rape and murder threats to be tweeted at women, including a female politician. Twitter doesn’t allow threats but relies on its community to ﬂag them for removal and report them to the police. While Twitter has automated systems to weed out spam, tweets about direct violence and suicide require manual review. “Context matters,” says Harvey. “‘Hey bitch can be a greeting or form of abuse.”
Harvey professes to be an optimist despite it all. “There are bad things out there, but I work alongside so many people trying to stop it.” On her left wrist is a tattoo of the heart emoticon <3. “It means hope.”