Engineered-in-Edmonton Twitter bot combats misogyny on the campaign trail

A social media bot is attempting to combat misogyny in on the federal campaign trail, one tweet at a time.

ParityBOT is not a troll but an inspirational cyborg

Unlike other trolling Twitter bots, ParityBOT was engineered to combat sexism with messages of positivity. (Getty Images)

Artificial intelligence programmers are attempting to combat misogyny on the federal campaign trail with a very positive Twitter robot.

ParityBOT is using artificial intelligence to identify hateful or abusive tweets directed at female candidates. In response, it spits out a positive tweet from the ParityBOT account.

Twitter bots are automated accounts run by software instead of human beings. They are programmed to automatically tweet, retweet, like and follow other Twitter accounts.

While many are designed to maliciously spam and troll users' timelines, ParityBOT has been engineered to send out empowering messages like, "You are stronger than you know. Keep fighting." 

The software, which relies heavily on language filters, was engineered by Edmonton-based ParityYEG in collaboration with the Alberta Machine Intelligence Institute and artificial intelligence scientist Kory Mathewson.

The non-human software could make the political sphere more humane for female candidates, one tweet at a time, said Alexandra Hryciw, a spokesperson for ParityYEG, an initiative created to empower more women to enter politics. 

"We got together a list of all women candidates running federally and then plugged them into the artificial intelligence bot," Hryciw said in an interview Monday with CBC Radio's Edmonton AM. "It monitors all their accounts."

None of the tweets mention candidates and they don't single out any users. The supportive tweets are crowd-sourced from public submissions from across the country.

The bot will continue its campaign of positivity until election day.

"We, as a group, decided that we were going to tackle misogynistic, sexist, harmful, racist and hurtful tweets directed at women candidates," Hryciw said. "We're trying to raise the level of public discourse."

Hryciw said the project was inspired by the litany of threats faced by Rachel Notley during her campaign and her time in the premier's office. Notley received at least 11 death threats, plus hate mail, vulgar photographs, unusual phone calls and a macabre incident involving a pre-arranged funeral application. 

Female politicians are subjected to disproportionate levels of harassment, said Hryciw. Too often the vitriol discourages them from pursuing their political aspirations.

"The No. 1 barrier is this hateful social media and the domino effect that hateful social media has on their lives," she said. 

"It's not only directed at them, it affects their campaign teams, it affects the people managing their social media sites and it also affects their families." 

The first iteration of the project — launched during the Alberta provincial election earlier this year — was soon overwhelmed by abusive messages, Hryciw said. 

The program processed more than 12,000 tweets during the provincial writ period, tracking 54 nominated candidates from March 19 until April 16. It sent out 973 positive tweets through the writ period but it could have sent out more, Hryciw said. 

The bot detected so many abusive messages the team had to modify its algorithm and put a cap on the number of positive messages it could send, Hryciw said. 

"It was so overwhelming.  There was so much negative content coming in that the bot was picking up that we just had to say, we're not going to be able to respond to every single one."