When sitting down with Dr Kathryn Hempstalk (Dr Kat!), we asked, very deliberately, “explain what you do like we’re five.” She laughed and said, “I teach computers to solve problems that are painful for humans.”

Dr Kat, Auror’s resident AI expert, has had an incredible career to date, commercializing AI and increasing productivity across a variety of industries, including dairy farming, accounting, online marketplaces, and now in retail crime at Auror. She’s customer-driven, mentors with empathy and understanding, and knows that trust leads the way when it comes to effective and responsible AI.

Understanding the problem

Dr Kat, like a lot of kids, had a computer growing up. But ask her about the one in her home, and she’ll tell you bluntly: it was no fun. There were no games, it wasn’t a cool computer, but there was a book on how to program it. In the absence of good games, she saw the computer as a puzzle she needed to solve; her parents were on board - it meant Dr Kat was busy, and it meant she was (mostly) learning.

This is the thing with Dr Kat, she’s constantly figuring stuff out, putting the puzzle pieces together and making sense of it - even if it’s not the most alluring topic. When Dr Kat started her career, it was working with AI in dairy farming. She trained a computer to check if a patch was on a cow, and the computer confirmed if the patch was there, and if so, whether it had changed color. The color change would signal if the cow was on heat, and ready to be artificially inseminated.

This AI process would give a dairy farmer another four to six hours in their day - because they didn’t have to stand and look at the cows. The machine could do it far more accurately, and the end result also meant the farmer’s cows were impregnated in a far safer manner than traditionally. Interestingly, and something that’s common across industries, the application of AI here was about 5% of her job. The other 95% was, as she puts it, “learning about cows, and how you can prevent them from doing stupid things, just so the camera could take a good photo.” 

That’s the real key to AI, according to Dr Kat. It’s understanding the problem, understanding the nuance, and understanding the environment in which the problem operates. Dr Kat, with a wealth of experience behind her, knows that this is something juniors in the field face - they just want to work on AI, but the truth is, it’s not the solution they should be thinking about first, it’s about the problem and where it lies. This is the kind of thinking she’s bringing to Auror and we’re here for it.

At Auror

One of the first cabs off the rank for Dr Kat is suggested merges, a project we’ve been working hard on improving for some time, and now it’s getting supercharged with Dr Kat here. It is certainly Auror’s most hotly requested feature - and it’s not been an easy nut to crack. For those not in the know, when an event or incident occurs, retail frontline workers enter it into Auror and our platform automatically creates entities (profiles) for any people or vehicles involved. Auror, unlike traditional case management systems, has always supported matching different incidents or events that involve the same person/vehicle to connect the dots.

Now, though, Dr Kat is working on the next level of this, where Auror can suggest to its customers whether to merge countless matching profiles so all the information is in one spot. Unfortunately, there is so much crime occurring that it’s just not possible for humans to keep up with matching this data across thousands of offenders, so AI is necessary to drive more efficiency and accuracy.  Because of the clean and structured data captured by the Auror Platform, Dr Kat explains that our system will produce a higher quality and quantity of matches compared to more traditional systems when we begin to roll out machine learning in practice.

AI requires a safe pair of hands, and we have that in Dr Kat. She’s honest to a fault, she says, and believes that has served her well in this industry. “When I look at a problem and how we might solve it with AI, I ask myself ‘does solving this do a disservice to the AI community and the use of AI?’” Over the years, Dr Kat has seen plenty of AI that works on a technical level, but isn’t quite there on a moral level: “It’s pretty cool I can do this thing, but what’s the harm in what I’m doing? Even though some of the ‘harm’ might just be inconvenient or annoying, because it can be easy to automate things, some people just don’t consider the harm in that.”

AI is an incredibly useful tool, and sometimes potential damage is traded off for the value it has. With Dr Kat and the wider team, we will always question how we can add the most value and minimize potential harm - and that’s the responsible AI approach we’ll continue to take here at Auror.

The future with Dr Kat

Dr Kat is excited about the responsible AI potential at Auror. For her, it’s not just about the product, but about the wider community and its overall potential to prevent crime. Dr Kat enjoys helping people; she’s motivated by service and it comes through loud and clear. She’s curious, constantly wants to solve interesting problems in effective ways, and add a touch of Auror magic while she does it.

We’re so happy to have Dr Kat on board. Please feel free to reach out to her about anything (except dairy farming) here. She’s always up for a conversation on responsible AI, and how that can help our wider community.

Posted 
April 6, 2022
 in 
Inspiration
 category

Join Our Newsletter and Get the Latest
Posts to Your Inbox

No spam ever. Read our Privacy Policy
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.