Areto Labs
Published in

Areto Labs

Garbage In, Gold Out: Building Twitter Bots for Gender Equality

Written by Lana Cuthbertson, CEO & Founder of Areto Labs

In 2017, I founded ParityYEG, a nonprofit organization dedicated to getting more women to run for public office, alongside some amazing women, after more than a decade of volunteering and advocating for gender equality in politics. While we all came from vastly different backgrounds and political ideologies, we had one thing in common — we wanted more women to run for office.

Women make up about a quarter of elected officials, across all levels of government, across North America. So we set out to recruit, encourage, and support women to run for office, with a goal of gender parity — 50 per cent. We’d meet women for coffee to chat about their campaigns and while they all had a variety of concerns, without fail the nastiness of ‘campaigning online while female’ was a universal barrier to running. We knew women are subjected to a disproportionate amount of toxic hateful messages on social media, and gender based violence is very much a persistent and global problem, but how would we solve it, or at the very least encourage others to support those on the receiving end of it?

That’s how we came up with @ParityBOT.

In December 2018, I traveled to Montreal for the NeurIPS (Neural Information Processing Systems) conference for my day job at a bank. This conference is the largest conference in the world for artificial intelligence research.

I was there to present a paper I had collaborated on with my colleagues in the workshop on challenges and opportunities for AI in financial services. But I also had the opportunity to check out other workshops and sessions. I was thrilled to find the workshop on AI for social good.

One poster in the workshop caught my eye. It was about a data collection and training project on how often women in politics and journalism experience abusive and problematic tweets on Twitter. It was the result of a collaboration between Element AI and Amnesty International. They called their study Troll Patrol.

I was inspired by this work because after many years of volunteering to elect more women, I was eager to dive in and think about how I could help solve one of the major barriers women face when they consider running for public office: online abuse.

Here is the most important lesson of this article: take vacations. Because while I was relaxing by the pool in Mexico a few weeks later, I had an idea: what if we could take this idea and build a Twitter bot that could detect abusive and problematic tweets directed towards women candidates during a certain election, or in a certain jurisdiction, and then send out a positive tweet (or “positivitweet” as my friend at SAM desk said) for every toxic tweet it detected?

So we did it. I found an amazing collaborator who is in the know about machine learning methods and who was willing to help us build the bot. We deployed it initially during Alberta’s spring 2019 provincial election. This news story does a great job of covering the system and showing what it can do, and it features my collaborator and I explaining it.

Our next step was to gather feedback, refine the bot and deploy it again during Canada’s fall 2019 federal election.

@ParityBOT uses artificial intelligence to detect toxic tweets sent to women in politics which then posts positive tweets in response. @ParityBOT was deployed during two Canadian elections in 2019, with each iteration running for the duration of an election campaign, or 45 days. During this time more than 245, 000 tweets were processed and 393 candidates were tracked. But the best part…more than 20,000 positive tweets were sent and more than 400,000 impressions were made to encourage women and spread awareness about the issue of online abuse that women face when running for office.

In addition to analyzing social media metrics, we measured the effectiveness of this intervention through qualitative research with women candidates after both elections. Women were overwhelmingly positive about this idea, and reported that this type of support is exactly what’s needed to encourage women to overcome the barrier of online abuse in order to run for political office.

We presented a research paper and poster on this work at the world’s top artificial intelligence conference, NeurIPS 2019, in December in Vancouver, Canada, at the AI for Social Good workshop — the same workshop that had sparked this idea the year before.

It was clear this idea had legs. How could we use the power of this technology to impact other social issues we cared about? Much like domestic violence, online gender based violence can be invisible to you and I. If you’re not seeking it out, you’d never know it exists. But it does. We are now building more bots like @ParityBOT that detect this abuse and tweet about it in a safe and effective way to bring this issue to light.

We knew we could make a positive impact and harness the power of this technology to fight inequalities around the world, and so Areto Labs was founded based on the premise that we would develop technology to build humanity and spread truth, equality and positivity. Are we nuts? Maybe. Are we eternally optimistic? You bet. But my experience over the past 10 years has taught me a few things. Resources are scarce but passion is not. This organization, started by only a handful of women with a vision, has already made an impact. Areto Labs was the result of deeply understanding a problem (not enough women in politics), getting to the root of a barrier to entry (online abuse) and tackling the problem in a meaningful, measurable and impactful way, according to the people we were trying to reach. When people come together to push boundaries and try new things, anything is possible.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Areto Labs

Areto Labs

Areto Labs is building the future of work by using machine learning and behavioral science to make digital communities more positive and inclusive.