Three years ago, companies were wary of artificial intelligence because it had not yet made the transition from cutting-edge technology to everyday implementation. Then the COVID pandemic brought home the importance of digital transformation, and everything changed.
Gaining value from data insights became a high priority, and businesses saw the power that AI could bring.
“Organizations are really thirsty to figure out how do we actually add customer value,” said Vidya Setlur (pictured), head of research at Tableau Software LLC. “How do we actually build products where AI can move from a simple, cute proof of concept working in a lab to actual production?”
Setlur spoke with Lisa Martin, host of theCUBE, SiliconANGLE Media’s livestreaming studio, in advance of the Women in Data Science (WiDS) Worldwide Conference 2022. They discussed enterprise adoption of AI and what makes intelligent visual analytics tools really intelligent.
Organizations using AI must be data literate
Adoption of AI continues to increase as companies accelerate their digital journey. But as they search for ways to create value through AI, businesses have to be aware of the dangers of irresponsible data use, Setlur pointed out.
AI has become commonplace in society, simplifying previously tedious tasks, such as mapping a route, adding filters to an image, and language translation. But while AI has proved itself in taking care of repetitive tasks, it’s not always accurate. Setlur uses machine translation as an example.
“You open up a Google page in Spanish, and you can hit auto-translate,” she said. “And it will convert it into English. Now, is it perfect? No. But is it good enough? Yes.”
But good enough isn’t acceptable if the AI is set loose to make decisions that affect business operations, and people’s lives, without any oversight.
As organizations build AI into their applications, they must become data literate, according to Setlur. This means making data a first-class citizen within the company culture and understanding the data that feeds into the algorithms. Looking for hidden biases and acknowledging where there are limitations in the data is important, as is being completely transparent about data practices and sharing them with customers. This can only happen if the human maintains control over the AI, according to Setlur.
“If these experiences are designed where humans are, in fact, in the driver’s seat … they can intervene and correct and repair the system if they do see certain types of oddities that come into play with these algorithms,” she said.
Show your support for our mission by joining our Cube Club and Cube Event Community of experts. Join the community that includes Amazon Web Services and Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger and many more luminaries and experts.