Consumers hold companies responsible for AI chatbot errors

From travel bookings to legal advice and even the political sphere, a growing concern surrounds the accuracy of information delivered by AI chatbots. One recent example involved an airline’s AI chatbot that promised a discount to a passenger booking a flight for a family emergency, only to reveal later that the discount wasn’t available. This incident and others like it have sparked a conversation about accountability when AI chatbots provide misleading information.

In law and policy, concerns have arisen about chatbots offering inaccurate legal answers or city policy details. The political landscape further complicates matters, with reports of AI chatbots generating false information about US elections and voting procedures.

To understand consumer sentiment on this issue, YouGov surveyed over 18,000 consumers across 17 markets. The survey delved into who should be most to blame when AI chatbots provide inaccurate information. Interestingly, a majority (54%) of respondents placed the responsibility primarily with "the company using the chatbot." Only 26% felt the onus fell mostly on "the developer of the company's chatbot”. A small number (3%) believed the customers were most responsible, and 3% felt no one was to blame.

This finding suggests that consumers hold companies accountable for the information disseminated through their AI chatbots. They view the company as ultimately responsible for the technology it deploys and the information it provides to customers.

When asked about culpability in general, here’s a breakdown of consumer opinion on company responsibility for misinformation delivered through chatbots:

• Globally, 71% of consumers believe that companies should generally be held responsible if incorrect information was provided on their behalf via the company’s AI chatbots.
• Consumers in several countries felt strongly about this, with Great Britain (81%), Australia (80%), Hong Kong (78%), and Canada (76%) leading the pack.
• French (57%), Indian (63%), and German (64%) consumers were less likely to place the blame with companies.
• In the US, 74% believe companies are responsible for AI chatbot errors, aligning with 76% for customer service mistakes and 79% for website misinformation.

The YouGov survey also explored whether companies should be liable for misinformation delivered through other channels alongside AI chatbots. Across all surveyed markets, four-fifths (79%) of respondents (ranging from 67% in France to 88% in Great Britain) felt companies should be held responsible for incorrect information provided on their websites. Similarly, a strong majority (76%) believed companies should be accountable for errors made by customer service representatives.

These findings demonstrate a clear trend: consumers expect companies to take ownership of the information they provide, regardless of the channel used. This includes ensuring the accuracy of information delivered through AI chatbots, company websites, and human customer service representatives.