Science and TechArtificial Intelligence

Actions

How accurate and effective are chatbots, and when should businesses avoid using them?

An expert joined Scripps News' "Capitalizing Technology" series to share the risks associated with American businesses increasingly using chatbots.
A person's phone shows ChatGPT.
Posted

The increased incorporation of artificial intelligence into our daily lives has manifested in different ways for different demographics.

For the average person, it could be with their digital voice assistants, like Alexa or Siri, or through listening to an app's music playlist recommendations. Meanwhile, the healthcare industry may be using robots to assist in surgeries, while social media companies can use it to create images.

Then there's the business sector, leaning into AI to handle tasks associated with cybersecurity, recruitment, and advertising, as well as customer service through the use of chatbots.

Chatbots are computer programs designed to simulate conversation with humans online. While ChatGPT is the most popular of them, Google, Microsoft and other developers have also begun offering similar programs. And businesses have welcomed them.

RELATED STORY | Here's how businesses are using artificial intelligence in hiring and how it could lead to bias

Data shows that 39% of all chats between companies and consumers involve a chatbot, according to reports by Dashly and Master of Code. But how accurate and effective are they?

Levi Brooks — CEO and co-founder of Use All 5, a design and technology agency — told Scripps News that it varies by model.

He said that although the big-name brand programs work well, they may struggle with recent information, complex topics, or conversations that mix subjects. There can also be bias or inaccuracies, which is why Brooks recommends always verifying the information a person gets directly from a chatbot.

Another concern with chatbots is privacy, particularly in the medical and financial sectors. Although businesses must comply with data protection laws, Brooks says there's always a risk involved with sensitive information being transferred through a bot.

"If I was a business, if they're dealing with any information that's very sensitive, I would keep it with the human at this point," Brooks said. "It's not worth it, kind of, commingling AI chatbots with very sensitive personal, private information."

Brooks also shared how businesses have shifted their use of chatbots, which sectors they're most used in, and if there are safety concerns with communicating general information through a chatbot.

To view Brooks' full interview with Scripps News, watch the video above.