Sign In

Communications of the ACM

ACM TechNews

AI Becoming More Conversational. But Will It Get More Honest?

View as: Print Mobile App Share:

With the Character.AI chatbot, users can speak with virtually anyone, from William Shakespeare to T'Challa or Queen Elizabeth II. "Remember," the site cautions, "Everything characters say is made up!"

Credit: The New York Times

Many companies and laboratories are tapping artificial intelligence (AI) to develop more sophisticated chatbots.

For example, the Character.AI website allows users to converse with any real or imaginary virtual character, including avatars of living and deceased individuals.

These bots often produce untruths, hate speech, and language prejudiced against women and minorities because they are trained on data posted online by actual people.

Some experts worry that the chatbots will mirror toxic biases and data without safeguards, while companies like Character.AI expect the public will accept their shortcomings and adopt a healthy distrust of their discourse.

However, experts concur that the technology's conversational ability will continue to advance, especially as people continue to generate vast amounts of text.

From The New York Times
View Full Article - May Require Paid Subscription


Abstracts Copyright © 2023 SmithBucklin, Washington, DC, USA


No entries found