Assisted Versus Augmented Intelligence
Produced by Steve Nathans-Kelly
Artificial intelligence, or AI, is not mature enough yet to provide full autonomous decision support, says Jeremy Bentley, CEO and founder of Smartlogic, who discussed the current state, benefits, and limitations of AI in his presentation at KMWorld Connect 2021.
Actually, said Bentley, where we're really looking at the benefits of any kind of AI, particularly in semantics or in some kind of AI analytics, or in the basic robotic automation of processes, we're talking about—as we always have with software—about automating the next level of repetitive work because it's compelling from a cost and efficiency perspective.
Now when we are automating the next level of work, and in many cases, we're talking about the automation of the extraction of data out of text, or the extraction of text-based data out of databases. The old way of working would have been taking 50,000 contracts to analyze, because there is some regulator coming to inspect them, and sending them to an army of a thousand paralegals to go and look at those contracts and pull out the data, he noted.
"Well, that's the kind of thing I'm referring to as the next level of repetitive work." It's automating the "stare and compare"; it is automating the extraction of data out of text, which of course is locked up in the content. You can't ask a PDF question.
"But when you're doing this, at the same time, there's a data quality yield, which is the consistent application of digital-driven processes to give accurate and traceable outcomes," said Bentley. "And this data veracity is important if the fidelity of your business decisions is important to you." You need to have all of the data and not just the data sitting in databases, but data that is internal and external to your enterprise, data that is structured and unstructured, data sitting in silos.
"If it's important that you've got the right set of data to be able to reveal better decisions, then data veracity is important," said Bentley. "And it is semantic AI that supports the automation of these manual processes by extracting that data and facts. And it's semantic AI, that's improving the veracity because we're able to harmonize and enrich information assets, be they structured or unstructured, or internal and external, to the organization."
Save the Date for KMWorld 2022—November 7–10, 2022—JW Marriott | Washington, DC!