AI and BI at UNSW

This week we're joined by Kate Carruthers, Chief Data and Insights Officer at UNSW, a university in Sydney with nearly 80,000 students and researchers. Kate talks about the interesting evolution of her role and her team's, from producing reports to becoming the data engineers and insights. We start by talking about data, and as we move to talking about Artificial Intelligence, Kate points out that "AI is the thing that is not yet in production", because once it's in production it's called something else (exactly as we found out when we talked to Troy Waller on the podcast, about the use of AI for accessibility, where AI wasn't the thing to focus on, it was the service it provided - like captions, dictation, text to speech etc).

Kate talks about the way to tackle AI and data problems - start with the problems of the organisation, not with the technology, and as she points out "If you start with the technology in mind, then you end up shaping the problem to fit the technology".

Kate also talks about UNSW's clear model for ownership of the data in the university - this is an important discussion, because in many cases using AI requires good organisational data, and in larger organisations it can be tricky to track down the data, and identify who can give permission for it to be used. In fact, in most AI projects, sorting out access to the data, accessing it, and tidying it up makes up 80% of more of the time and effort!. And as Kate makes clear, this isn't just about how the university uses the data, it's also about clarity on how student data cannot be used!

Finally, Kate discusses ethics and governance of data and artificial intelligence, and the work that is being done in the university to build a policy for AI use, alongside the existing clear policies on data.

 

In the podcast, Ray talks about the large proportion of AI projects that fail to deliver business benefit, and he conservatively talks about 60%. But there's a number of published reports and articles that put the number much higher:

• July 2019: VentureBeat AI reports 87% of data science projects never make it into production
• Jan 2019: NewVantage survey reports 77% of businesses report that "business adoption" of big data and AI initiatives continues to represent a big challenge for business. That means 3/4 of the software being built is apparently collecting dust. Ouch.
• Jan 2019: Gartner says 80% of analytics insights will not deliver business outcomes through 2022 and 80% of AI projects will “remain alchemy, run by wizards” through 2020.