How do you go from being a police officer to one of Australia’s most recognised AI entrepreneurs? It’s not a traditional pathway, but it was during her time in the NSW Police Force that Dr Catriona Wallace first discovered a passion for organisational behaviour.
“I saw how the systems worked within the Police Force and how the Police Force worked as a system within society. It was pretty dysfunctional, and I could see why it was. I wondered if technology could improve this,” says Wallace.
Wallace left the Police Force in 1990 to study organisational behaviour. She predicted that organisations would become reliant on technology to make certain decisions. Her PhD examined the ways that computers could be used as a substitute for human leadership. “It was the first study of its kind that looked at when technology would actually replace, enhance or detract from the effect of the human leader,” she says.
After completing her PhD in 2007, Wallace co-founded a market research firm called ACA Research, a customer experience consultancy called Fifth Quadrant and she also founded a Sydney-based co-working space called The Ventura.
Her mind kept turning back to how organisations could use AI to improve how they functioned. She oversaw the building of a platform that would help financial services companies improve online sales conversion rates and customer service. In 2014, she launched Flamingo AI, which developed AI-powered chatbots for the financial services sector. In 2016, it became the second woman-led business ever to be listed on the Australian Stock Exchange. Flamingo AI was sold to private investors in 2020.
Wallace has seen radical changes in start-up funding in Australia.
“Seven years ago, there was hardly a start-up community. We didn’t use the word ‘entrepreneur’,” she says. “There were no incubators or start-up accelerator programs. While I founded Flamingo AI in Australia, within six months, I took it to the United States, and we built the company from New York. Then, in 2016, when things had picked up in the Australian market, I brought it back on shore and listed the company.”
When she was initially looking to raise funds for Flamingo AI, Wallace says there was only $20 million invested in start-ups in Australia. “Now it’s billions.”
Raising capital as a woman is a uniquely difficult task.
I reckon I work 150% harder than my male equivalents in order to establish credibility and to be taken seriously as a founder, a CEO, and a technologist,” says Wallace. “I often get a lot of mansplaining – investors, or people in the industry, trying to tell me what artificial intelligence is, or how to run a start-up. That’s where it’s been very interesting having a PhD. I would usually introduce myself as Catriona Wallace, but in front of investors and the like, I will put the ‘Dr’ in there.”
“Then I know I’m dealing with either conscious or unconscious bias. For women, it’s particularly difficult to raise capital in Australia. It’s difficult the world over. About 2% of the world’s capital or venture capital goes into women-led businesses, so we’re still a long way from being anywhere near equal. But it’s certainly improved.”
Wallace sees the lack of women in the capital markets – either as investors or in leadership positions within funding organisations – as one of the biggest issues holding women founders back. “I call it a ‘hyper-masculine’ environment. They’re not used to women in leadership positions,” she says. “They don’t have the experience to be dealing with us. I also believe that the way women present, the things we care about, are different to what our male counterparts care about and how they would present. I think that gets lost in translation.”
Wallace advises that there is some advantage in the rule of thumb that if there is a woman presenting, then there should be a woman in the room to help amplify what she is saying. “If there’s a woman founder or CEO who’s presenting, and perhaps she’s not presenting in the formulaic way that investment conversations usually happen, then there should be a female in the room to say, ‘Right, I understand. I hear what you’re saying. You meant this when you said that? Great. Yes, we understand that. Did you say this? Yes, great. That’s what you meant, excellent. Continue on.’”
Wallace also recommends that aspiring women entrepreneurs attend ‘pitch bootcamps’, such as the invite-only Springboard Enterprises Accelerator Program in the US, which is designed to accelerate the growth of entrepreneurial companies led by women. Wallace believes this type of training is essential and worth the effort of applying. “We need more of these bootcamps, so women can be trained in how to present and operate in this capital-raising world, because it is tough,” she says. Wallace also recommends for those in the AI field to consider accelerators such as Boab AI, which encourages women entrepreneurs and delivers programs to help young companies scale.
Laying the foundations for ethical AI
The thread that has always held Wallace’s work together is her focus on how we remove society’s unconscious bias from AI and how we ensure that the AI we create is ethical. Wallace points to two issues – a lack of diversity in those who are funding and developing AI technologies, and an established ethical framework that governs all business.
“While I was working in the Flamingo AI business in the US and Australia, it occurred to me many, many times that the people who were using or developing AI had no understanding of ethics or how to use the technology responsibly,” she says. “It didn’t mean that they were using it irresponsibly, but there was just no knowledge about the potential danger of this technology, particularly in algorithmic decision-making.”
In recent years, Wallace has become increasingly focused on ethical AI and responsible AI frameworks. In 2019, she acted as pro-bono advisor to the then Australian Minister for Industry, Science and Technology, Karen Andrews, on Australia’s AI roadmap and Australia’s Artificial Intelligence Ethics Framework.
In 2020, Wallace launched a company, Ethical AI Advisory, which provides advice to government, corporates and start-ups on how to develop responsible and ethical technology. She is also working with the Gradient Institute, a non-profit Australian research institute that works to build ethics, accountability and transparency into AI systems. “While I do more of the governance at the board level and executive level, Gradient works with the engineers and the data scientists training them on ethics,” she says.
Together, Ethical AI Advisory and the Gradient Institute are about to release their first big report based on a study of more than 400 Australia-based organisations on their usage of, or understanding of, responsible AI. “We really want to try and step up Australia’s understanding and use of ethical technology,” says Wallace. “The difficult thing is, on an international basis, we’re so far behind in AI. The US invests 10 times more per capita than we do.”
Where ethics and commercial interest collide
In 2020, Google made the controversial decision to oust Timnit Gebru, technical co-lead of its Ethical Artificial Intelligence Team, following the publication of a study she led that was critical of large language models – a type of AI software that powers Google’s search engine. Many researchers in AI responded by asking, how can ethical use of AI be governed in big tech companies if it’s not in their commercial interest?
Big Tech are trying to train their people to do ethics, but the sacking of Timnit Gebru is an example of where we are at,” says Wallace. “The film The Social Dilemma really captured it. Big tech is where a lot of the problem is. I think that’s because their business models are not properly aligned with ethics. It doesn’t mean that they’re being unethical or irresponsible – that’s important to acknowledge – but the way their business models and their profit is generated is on using algorithms and AI to manipulate populations in certain ways to buy things, and to vote in certain ways, for example.”
Wallace points to representation as a major issue that is holding back reform in the technology industry. “There are still hardly any women at the table. It’s one in 10 in artificial intelligence, and fewer than that in leadership. It’s still predominantly white male leadership, and at this stage, ethics doesn’t feature highly on the transformation agenda.”
Accelerating diversity
Wallace has recently stepped into the role as Chair of Boab AI, Australia’s first AI Scale-Up Accelerator, created in partnership with Launch Vic and backed by Artesian Capital.
Boab AI intends to fund 32 scale-up AI companies in Australia, with 50% of them based in Victoria, over the next four years. The support includes funding and a six-month program to help fast-track them through their scale-up phase. Boab AI is also raising a $100 million fund so it can provide funding and support for another 150 companies globally.
Wallace and her team are working with universities to get their researchers to present to Boab AI, which can help them determine whether there is a commercialisation avenue for their ideas.
She also wants to build diversity.
We want to find that nerdy, transgender kid, who’s sitting in their room, creating amazing technology, but who may not be confident or think anyone might invest in them at all. We want to go to Indigenous community, the disabled community, the women’s community, the refugee community. We want to support women and minorities who would normally fall out of the funnel somewhere,” says Wallace. “I want to change what an AI entrepreneur looks like.”
Article by Kylie Ahern
Photo Credit: Photo supplied