Winner Profile: How Jacqueline Comer is using AI to make the online world safe enough for everyone to show up

Jacqueline Comer, based in Auckland, New Zealand, is the Founder and Chief Product Officer of Areto Labs, a Canada-based company building AI systems that detect and remove online harm in real time. She won the She Shapes AI Global Award in the AI + Safety category at the 2025/26 She Shapes AI Conference and Global Awards Ceremony, held on 16 April 2026 at the London School of Economics.

The numbers behind her work are striking. Since August 2022, Areto's systems have analysed nearly 24 million online interactions in real time, identified and removed over two million abusive or harmful exchanges, and counteracted more than 10,000 gendered microaggressions. Across communities using the platform, hate-speech harm has been reduced by up to 99 percent and microaggressions by 98 percent. In the past year alone, the system intercepted 229,000 fraud attempts, disrupted 34,000 bots, and removed 150,000 harmful interactions. Areto's technology has protected athletes during the Women's Cricket and Rugby World Cups, and its Global Hate Speech Index for Sport is establishing new international standards for digital safety and accountability.

Jacqueline's path to this work began not in a computer science lab but in linguistics and writing. An English major who became fascinated by the impact of words, she arrived at AI through a question most systems had failed to answer: why were gender microaggressions slipping through? When she started looking, she found bias baked in at every level of the pipeline: the data, the annotators, the people deciding what accuracy meant. That discovery shaped everything that followed.

On what it takes to build AI that genuinely protects people online, Jacqueline is direct, grounded, and for someone doing some of the most serious work in the field, refreshingly undramatic.

Being part of the hopeful conversation

When Jacqueline received the email telling her she had won, her first thought was: "I can't wait to tell my mum." That instinct says something about how she holds her work: seriously, but without pretension.

"There’s so much bad news about AI right now. So it feels like a really nice time to showcase some really great stuff going on in the world of AI. It's really great to be able to be part of the hopeful conversation rather than the doom and gloom conversation."

After six or seven years working in AI and natural language processing, she describes recognition like this as a moment of alignment: a chance to demonstrate, with evidence, that the technology producing headlines about harm is also the technology being used to counter it.

The cost of accepting online harm as normal

The problem at the core of Areto's mission is one that has become dangerously normalised: the assumption that online abuse is simply the price of being visible.

"We're asking certain segments of society, whether you're a social media manager or any sort of public-facing person, to just accept that being attacked online is part of the job. And that's wild. That's a terrible way to exist."

The consequences extend beyond individual distress. Online abuse shapes who steps forward for political leadership, who feels safe building community in public, and who takes on roles that require any kind of leadership and online presence. The impact falls acutely on women, LGBTQ+ communities, and people from equity-seeking groups: precisely those whose voices have been historically, and now systematically excluded from leadership and the conversations a functioning society depends on.

The scale of the problem has also outpaced human capacity to address it. Areto has documented a client receiving one spam comment every second for two hours during a single football match. "It's almost like a joke, robot versus robot, because the volume of what's coming in, you just can't keep up." And there is another reason to want a machine doing this work: reading abusive content at volume causes serious harm to the people doing the reading. Most people who moderate content are not hired as content moderators. They are hired to post, write, and manage. AI takes that burden away.

The moment the bias became impossible to ignore

Jacqueline's defining moment came early, when she was analysing data from a prototype Twitter bot designed to respond positively whenever women election candidates were targeted with toxic content.

"I started to look at what we missed, like why did the AI not catch this horrible thing? And when I started to analyse the data carefully, I realised our model was missing gender microaggressions."

When she traced why, the answer was in the pipeline itself: where the training data was sourced, who was annotating it, who was deciding what counted as accurate. Bias was not an edge case. It was structural, and it was overt. A political leader being called "climate Barbie" was not being caught. The implications were serious.

"If these systems that we all start to adopt have bias baked in, and it was not even subtle, then what damage are we going to do because of this?"

That question became the foundation of Areto's approach: diverse training data, diverse annotators, continuous bias monitoring, and models built to detect the nuanced, intersectional forms of abuse that generic systems routinely miss. There is also a commercial logic to this rigour. "Responsible AI and trying to remove bias actually makes better models. And making better models means you're doing better, more accurate work, which means you have a better business."

The good stories need a louder voice

Asked about the gap between how AI is discussed and how it is actually being applied, Jacqueline's answer is less technical than it is human.

"There's so much talk about all the crappy stuff about AI. But there's also hope, and I think that gets missed. There are so many problems that nobody is talking about that somebody is thinking about right now and trying to solve with AI. And I would love to tell more of those stories."

Fear, she notes, is persuasive. And when it drives people away from understanding and using AI, the knowledge gap tends to fall hardest on those already most affected by its harms. "The more fear we have about it, the less people are likely to use it, understand it, and that's just going to create a bigger problem."

On leadership in AI, she returns to the same idea: more leaders doing work that matters need to tell their stories, share what they know, and engage publicly with why they are building what they are building. Underpinning all of it is accountability, for the intended impacts of what you build and for the unintended ones. "The last thing you want is to use a tool to solve a problem and create more problems somewhere else."

What comes next

Areto Labs continues to extend its reach beyond sport, with customers across media and publishing and public service, including a research partnerships now underway with the Global Institute for Women's Leadership, applying its technology to protect women in office through its Face Forward initiative. The Global Hate Speech Index for Sport, meanwhile, is already being explored as a replicable framework for politics and public life.

For Jacqueline, the measure of progress is ultimately simple: more people feeling safe enough to show up online, take on public roles, and use their voices without paying a toll in harassment and harm. "The online world is open for everyone, which is great, but also has a lot of negative impacts." The work of Areto Labs is to close the gap between those two facts.

Jacqueline Comer is proof that building responsibly and building well are not in tension. They work towards the same goal.

Learn more about Areto Labs and their work at aretolabs.com.

The 2026/27 She Shapes AI Global Awards Call for Applications opens on 8 October 2026. If you or someone you know is driving positive impact through AI, sign up to our newsletter to stay informed.

She Shapes AI is the global catalyst for women-led, responsible AI innovation. Founded in 2024 by Dr Julia Stamm FRSA. sheshapes.ai