In late 2016, Bank of America announced the expected debut of its chatbot offering.
Chatbot supporters, however, argue that the technology is just as secure as the voice response systems customers use today.
“Most financial institutions have substantial volumes of content and it can be difficult to get the right information.”Finding the nearest ATM or branch (which some virtual assistants such as Siri can already do) “is not the most useful use” of chatbots, Oren says.
“You can get the same experience by using Google Maps.
As part of the learning process they set up two bots, known as a dialog agents, to teach each other about human speech using machine learning algorithms.
The bots were originally left alone to develop their conversational skills.
I rarely book an appointment online because it requires me to create a new account for their enterprise salon management platform.
“When the bank bot lives within the bank application, it’s just as secure as the bank’s Web site and servers,” says Dror Oren, Kasisto’s vice president of product.
Chatbots, he adds, “mirror the same security settings that banks already offer today,” At the very least, it is hoped that if security problems arise, early pilot programs will identify them—and let banks and tech companies find remedies before a rollout.“The use case where I could see chatbots working fairly well is in helping customers or prospects find the right pages for content,” Knieff says.
However, researchers at the Facebook AI Research Lab (FAIR) found that they had deviated from script and were inventing new phrases without any human input.
Facebook's Artificial Intelligence Researchers were teaching chatbots to make deals with one another using human language when they were left unsupervised and developed their own machine language spontaneously (stock image)Scientists and tech luminaries including Bill Gates have said that AI could lead to unforeseen consequences.