Google is Already Worried About Regulation of Machine Learning

Machine learning could be quite influential on cloud services (and FIs), which is why Google’s experiments with it are ongoing, according to Gavan Corr, principal, financial services for Google Cloud.

“Financial services has been at the forefront of machine learning — a lot of what we call machine learning is really math. However, not everyone is deploying machine learning right now, because we haven’t solved a few problems,” said Corr, pointing specifically to regulation.

Regulatory concerns are one of the prime reasons financial institutions have not yet integrated machine learning or related technologies into their main processes, Corr, whose background is in financial services (previous to Google, he worked with NYSE Technologies and Lucera), noted yesterday at the FinDEVr conference.

“It’s going to be very hard to tell a regulator that the computer made the decision, hand them a piece of math, and say ‘Isn’t it obvious?’” he said. “So, we really have to work with the regulators to educate them.”

In the meantime, Google is focused on growing its cloud services.

Corr cited the company’s familiarity with “large-scale distributed computing,” and its willingness to experiment with technologies like machine learning as some of the advantages of the company’s offering. But companies should come up with a project idea first, and only then pick a cloud service, he explained.

Amazon’s cloud offering–Amazon Web Services, or AWS — was the first platform to bring cloud infrastructure as a business “to the masses,” according to Corr. But there is no lack of usage for Google’s offering either, he added.

Google Cloud already supports seven unique products, said Corr — including Google Maps, the Play Store, and Gmail services — with 1 billion users.

“Amazon and Google are very stable platforms–they are targeted at very different markets. It has to fit with what you’re doing,” said Corr. “If scale and machine learning are an issue, I would pick Google.”