DPD’s AI-Powered Chatbot Disabled After Irate Customer Made It Swear, Criticize Company

Dynamic Parcel Distribution (DPD) is part of Geopost, an international delivery service for sorter-compatible parcels based in France. In a turn of events, DPD decided to turn off its AI-powered chatbot after it seemingly dissed the company to sympathize with the disgruntled customer.

DPD Disable AI-Powered Chatbot

Musician Ashley Beauchamp, 30, was attempting to locate a lost package, so he contacted DPD's chatbot for support. However, he got frustrated since it wasn't helpful at all. Due to his frustration, the irate customer switched gears and started playing around to see what the chatbot might accomplish. According to Beauchamp, this is when the "chaos started."

He first asked it to tell him a joke, but he quickly moved on to asking it to compose a poem that was critical of the business.

Beauchamp posted the exchange on X, to which the chatbot responded with the phrase, "F**k, I'll do my best to be as helpful as possible, even if it means swearing."

In another part of the conversation, the chatbot even refers to itself as a "useless Chatbot that can't help you."

A single post from London-based classical musician Beauchamp received 800,000 views in a single day. "It's utterly useless at answering any queries, and when asked, it happily produced a poem about how terrible they are as a company," he said on X, referring to the chatbot.

Beauchamp's post received several responses from netizens. One wondered if it was genuine, adding that if it was, then it was "absolutely hilarious." Beauchamp assured her that the conversation was real. Another said it was an "AI revolution."

Many netizens expressed doubt over the conversation, with several asking if the exchange was "real." Beauchamp doubled down, telling them he had just taken the screenshots from his phone. He explained that he was trying to track a lost parcel and was entertained by his convo with the AI chatbot.

DPD uses artificial intelligence (AI) in its online chat to assist human operators in answering questions. According to the company, the chatbot's strange behavior was caused by a recent update. As a result, the company turned off the chatbot, removed the offending component, and updated its system.

Beauchamp pointed out that these chatbots were meant to make our lives easier, but when used incorrectly, they usually make things worse and make the user feel more impersonal and frustrated.

As for his lost package, DPD had already contacted Beauchamp to "resolve his issue."

ChatGPT Provides More Empathetic Responses

A panel of doctors preferred ChatGPT's sensible and empathetic responses over actual physicians. ChatGPT is a chatbot developed by OpenAI and launched on Nov. 30, 2022.

In a recent study led by the University of California, the effectiveness of doctors and the AI chatbotChatGPT in responding to 195 medical inquiries were compared to determine which was more empathetic.

Researchers collected questions from the public "AskDocs" subreddit. One guy, for example, asked about the dangers of swallowing a toothpick. Another person asked if they would have a concussion if they hit their head on a metal rod. A moderator verified the qualifications of the healthcare professional who responded to questions on the subreddit. The researchers also ran the questions using ChatGPT to generate a response.

A group of medical professionals were asked which response, the doctor or the chatbot, was better. They did not, however, know whose answers came from.

Each instance was assessed by three different judges, yielding a total of 585 evaluations. The judges' ratings were then averaged.

In 79 percent of the cases, the judges preferred the chatbot responses over the medical responses due to their superior understanding and more empathetic language.

Check out more news and information on Technology in Science Times.