The “technology risk” that surrounded artificial intelligence and the companies relying on it is largely gone, according to Blackbird ventures, a leading Australian venture capital fund that controls $1.24 billion in capital. 

Blackbird cautions, however, the technology must be matched with adequate access to data to attract capital.

In August Blackbird raised $500 million for its fourth fund, the largest in Australian venture capital history, and is an early backer of local success stories Canva and SafteyCulture.

Rick Baker, co-founder at Blackbird Venture, says five years ago AI was thought of as a “deep tech” – promising but accompanied by a substantial risk. Baker tells Which-50 investors at the time were conscious of the “technology risk” and whether companies could actually build with AI in a scalable way.

Rick Baker, co-founder Blackbird Venture. Image: blackbird.vc

“I think now, as we see it today, that a lot of that technology risk is gone. Because of the buildup of platforms around AI and machine learning,” Baker said at an AWS media event in Sydney this week.

The real question for investors now, Baker says, is whether companies will have robust access to useful data.

“You need to do the training [with AI and ML models] and can you get labelled data to do the training at scale?”

It means investors are often less concerned with algorithms and more concerned with data access, something company founders pitching to Blackbird often underestimate.

Red flags

Data access is particularly challenging in fields where the emerging tech holds more promise, such as medical diagnostics – an area Blackbird has already placed several bets on AI. 

“Founders come with this sort of stars in their eyes about being able to get hold of the data,” Baker tells Which-50. “Particularly when it comes to things like medical diagnostics, because it’s still really hard to get hold of that data.”

Blackbird is looking for companies that can pair their technology with structured and well labelled data, Baker says, because typical data comes in an “unstructured heap of mess”.

“We’re not actually looking for things that are AI for the sake of AI.

“What we are looking for is really great products and things that you couldn’t do four or five years ago that you can do now, and that computers can do as well as humans, and take out error rates, can take out expenses and really democratize things that were only available to a few people in the past.”

Machine learning medical

As for what sectors might lend themselves to AI technology, Baker says, while early days, medical diagnostics looks particularly promising.

“We do think the way that disease [and] conditions are diagnosed is going to change,” Baker told Which-50.

“At the moment it is often highly skilled people who have had years of pattern matching experience looking at medical imaging looking at symptoms, and making a diagnosis. As we know that diagnosis is subject to error. And you’re looking for the most senior person you can find to make that decision.”

That’s where AI and machine learning can provide value, according to the Blackbird cofounder, who says it’s an area where the technology can outperform humans.

“Actually we think that machine learning algorithms can do it better than humans. They are built to do pattern matching and that’s what this is about. So we think that’s a really, really good space.”

LinkedIn
Previous post

Stone & Chalk adds a former Myer and CBA executives to board

Next post

Reinvent or perish, AWS CEO tells enterprises