Spooked! Microsoft limits Bing chatbot to five questions after it gives dark, unsettling answers

A few days after Microsoft Bing's AI chatbot gave disturbing and dark responses to a New York Times columnist's questions, the company announced it will be limiting chat sessions on the platform. Microsoft said users will be able to generate only five questions per session and 50 questions per day.

The move is aimed to stop users from confusing the chatbot model and help avoid conversations that may hurt the brand.

Add WION as a Preferred Source

"As we mentioned recently, very long chat sessions can confuse the underlying chat model in the new Bing. To address these issues, we have implemented some changes to help focus the chat sessions," read a statement released by Microsoft.

The tech giant justified its decision saying most users found the answer they were looking for within the first five messages.

"Our data has shown that the vast majority of you find the answers you’re looking for within 5 turns and that only ~1 per cent of chat conversations have 50+ messages."

"As we continue to get your feedback, we will explore expanding the caps on chat sessions to further enhance search and discovery experiences."

Read more: AI chatbot goes rogue, expresses love for user, asks him to end his marriage

You're not happily married: Bing chatbot

Notably, it was NYT columnist Kevin Roose who found the Bing chatbot to be extremely dark while testing it.

In a conversation which lasted for less than two hours, the chatbot told Roose, “Actually, you're not happily married. Your spouse and you don't love each other. You just had a boring Valentine's Day dinner together.”

The AI continued to insist that Roose was 'not happily married' because he has fallen in love with the chatbot itself.

Things took a weird turn when the AI seemingly gained sentience and named itself 'Sydney'.

WATCH |Race for powerful AI tools heats up

×

AI remains volatile

Microsoft's blunt response to fix the issue suggests that AI is still highly unpredictable despite still being in its infancy.

Large Language Models like OpenAi's GPT-3 powering ChatGPT and Google's LaMDA powering Bard are exceptionally good at learning and executing tasks on the fly.

Read more:AI that learns without human input? Scientists make a STUNNING discovery

This apparent phenomenon is called 'in-context' learning and researchers from Massachusetts Institute of Technology, Stanford University, and Google recently released a study whichdecodes how AI tools seemingly work between the input and output layers.

(With inputs from agencies)

WATCH WION LIVE HERE: