C3.ai Expands Data Lake

The data lake just got bigger.

Data lake, you say?

A few weeks ago, enterprise AI software company C3.ai created the COVID-19 Data Lake, a free resource that unifies multiple datasets, highlights links between the sets, and displays it all on a knowledge graph. Today, C3.ai announced it’s doubling the lake with 11 new COVID-19 datasets.

• The data comes from sources such as Johns Hopkins, the WHO, the NYT, Europe’s CDC, genome sequences, epidemiology reports, and academic journals.
• Amazon Web Services is donating free cloud computing to the effort.

C3.ai’s bread and butter is designing and operating “large-scale commercial industrial AI applications” for corporations, founder and CEO Tom Siebel told me. “This is what we do professionally, so it really wasn’t much [additional] work.”

Use cases

Researchers are using the data lake to:

• Predict when to scale manufacturing as countries reopen
• Examine disease spread across populations
• Develop pandemic strategies and response scenarios
• Build predictive models of mobility and infection rates
• Support contact tracing
• Understand mortality rates of COVID-19 patients with pre-existing conditions

Reminder: AI is no silver bullet. Take diagnostics, where AI tools have stumbled in transitioning from trials to clinical settings. Researchers recently examined 31 AI prediction models for coronavirus diagnosis and prognosis, and preliminary findings suggest most models “are at high risk of bias…[and] likely to be optimistic and misleading.”

But more clean, structured, and machine-readable COVID-19 data is flooding in by the day, Siebel said. Even if AI tools cannot accurately replace radiologists, they can quickly analyze population-level data. And AI’s modeling powers only get better with more data.

Bottom line: “As of March 14, 2020, I cannot think of a more important application for AI,” says Siebel. “This is a large, important global issue with enormously consequential social, economic, and public health implications that go way beyond COVID.”