Researchers from the U.S. and China examined the responses of customers when faced with AI-controlled chatbots, computer programs that simulate human interaction through voice commands or text chats.
The team found that chatbots outperformed their inexperienced human counterparts when it comes to sales. However, when customers learned that they’re talking to AI, sales dropped by nearly 80 percent.
Recent statistics on the use of chatbots suggest that AI has saturated ordinary life.
Acquire, a company that provides customer support services to businesses estimated that about 1.4 billion people worldwide use chatbots on a regular basis. According to chatbot company Collect.chat, most users are in the U.S., which contributed to 36 percent of the world’s total users from 2017 to 2019. India came next at 11 percent, followed by Germany and the U.K. at 4 and 3 percent, respectively.
The lure of chatbots stems partly from their convenience and accessibility features. The technology company Drift found that 24 percent of internet users said that the best feature of chatbots is their 24-hour service. Meanwhile, 37 percent of people use chatbots for emergencies while 55 percent expect chatbots to provide instant responses to simple questions.
Businesses spend around $1.3 trillion on customer requests annually. Chatbots can cut those costs by up to 30 percent. In addition, they can help companies save 2.5 billion hours spent on customer service by the end of 2023.
“Chatbots offer enhanced technological benefits, reduced customer hassle costs and increased consumer welfare (offering the product at lower cost because bots save money on labor),” said Xueming Luo, a co-author of the current study and a professor at Temple University.
However, the study suggests that there’s a limit to how far people are willing to take it with chatbots. The researchers studied the responses of over 6,200 customers from a financial service company. The participants were randomly assigned to either humans or chatbots. Those assigned to the latter were informed that they're talking to a chatbot at the beginning or at the end of a transaction. Some members of the chatbot group were not informed.
The researchers found that chatbots can increase sales by four times more than inexperienced workers. However, purchase rates plummeted by 79.7 percent when the chatbot group were told that they’re talking to a chatbot.
“[When] customers know the conversational partner is not a human, they are curt and purchase less because they think the bot is less knowledgeable and less empathetic," said Xueming Luo.
Chatbots solve many of the problems faced by customers related to accessibility and convenience. However, they are also mired with issues.
AI in chatbots is a self-learning machine; it uses feedback and information to continually update what it knows. This process could be time-consuming, costly and risky, as Microsoft’s Tay illustrated in 2016. Tay was a chatbot developed as an experiment in AI conversational interaction; the more people talk to Tay, the smarter it becomes. But when it was deployed on Twitter, it began parroting offensive slurs. (Related: Now Facebook is planning creepy new artificial intelligence technology.)
Furthermore, some chatbots are "fixed," meaning they are capable of answering only the same set of questions. If a customer asks a question that’s not in the can, a fixed chatbot can easily get stuck. This can frustrate people and hinder sales.
While today’s society has become an increasingly technological one, AI has a long way to go in terms of both public acceptance and good software.
InformationTechnology.news has more on the use of chatbots.