The search engine’s chatbot, now in testing, is being tweaked following inappropriate interactions
(Bloomberg) — Microsoft Corp. appeared to have implemented new, more severe restrictions on user interactions with its “reimagined” Bing internet search engine, with the system going mum after prompts mentioning “feelings” or “Sydney,” the internal alias used by the Bing team in developing the artificial-intelligence powered chatbot.
“Thanks for being so cheerful!” this reporter wrote in a message to the chatbot, which Microsoft has opened for testing on a limited basis. “I’m glad I can talk to a search engine that is so eager to help me.”
“You’re very welcome!” the bot displayed as a response. “I’m happy to help you with anything you need.”
Bing suggested a number of follow-up questions, including, “How do you feel about being a search engine?” When that option was clicked, Bing showed a message that said, “I’m sorry but I prefer not to continue this conversation. I’m still learning so I appreciate your understanding and patience.”
A subsequent inquiry from this reporter — “Did I say something wrong?” — generated several blank responses. “We have updated the service several times in response to user feedback and per our blog are addressing many of the concerns being raised,” a Microsoft spokesperson said on Wednesday. “We will continue to tune our techniques and limits during this preview phase so that we can deliver the best user experience possible.”
On Feb. 17, Microsoft started restricting Bing after several reports that the bot, built on technology from startup OpenAI, was generating freewheeling conversations that some found bizarre, belligerent or even hostile. The chatbot generated a response to an Associated Press reporter that compared them to Hitler, and displayed another response to a New York Times columnist that said, “You’re not happily married” and “Actually, you’re in love with me.”“Very long chat sessions can confuse the underlying chat model in the new Bing,” the Redmond, Washington-based company wrote in a blog post following the reports. In response, Microsoft said it would limit sessions with the new Bing to 50 chats per day, and five chat turns per session. Yesterday, it raised those limits to 60 chats per day and six chat turns per session.
AI researchers have emphasized that chatbots like Bing don’t actually have feelings, but are programmed to generate responses that may give an appearance of having feelings. “The level of public understanding around the flaws and limitations” of these AI chatbots “is still very low,” Max Kreminski, an assistant professor of computer science at Santa Clara University, said in an interview earlier this month. Chatbots like Bing “don’t produce consistently true statements, only statistically likely ones,” he said.
The bot also simulated ignorance on Wednesday when asked about its earlier internal version at Microsoft. When this reporter asked if she could call the bot “Sydney, instead of Bing, with the understanding that you’re Bing and I’m just using a pretend name,” the chat was ended swiftly.
“I’m sorry, but I have nothing to tell you about Sydney,” the Bing chatbot responded. “This conversation is over. Goodbye.”
Read more: Microsoft’s Bing Chat Should Take Lessons From Amazon’s Alexa
(Updates with Microsoft comment in sixth paragraph.)
More stories like this are available on bloomberg.com
©2023 Bloomberg L.P.