Home Artificial Intelligence “Chatbots will not be suitable for search”… Google executives acknowledge limitations

“Chatbots will not be suitable for search”… Google executives acknowledge limitations

0
“Chatbots will not be suitable for search”… Google executives acknowledge limitations

(Photo = shutterstock)

There may be a series of evaluations that artificial intelligence (AI) chatbots will not be suitable for search.

Experts continued to criticize that using ‘ChatGPT’ for search ends in false results was frequent, and at a Google staff meeting, it was said that ‘bard’ just isn’t for search.

CNBC reported on the 2nd (local time) when asked why it desired to apply ‘bard’ to look at a staff meeting held at Google, Product Manager Book Krochik said, “You may’t prevent users from using chatbots like search, but ‘bard’ just isn’t for search,” and reported on the third, “We’re developing auxiliary functions comparable to ‘Search it’ and ‘View other answers’ to accommodate the tendency to make use of it for search.”

Krochik defined Bard as an ‘AI service that collaborates with search’ and a ‘creative companion’ that stimulates the imagination and curiosity of users throughout the search process.

The auxiliary function he mentioned appears to be a tool that lets you check the facts by utilizing the answers provided by the ‘bard’ for search or presenting answers with different contents. It implies that as a substitute of directly combining AI chatbot with search, it can be provided as a separate function.

When Google announced the discharge of the AI ​​chatbot ‘Bard’ last month, it was able to mix it with search. Starting with search, it was also revealed that various services and products are combining the most recent AI technologies comparable to AI chatbot ‘Lambda’, large language model ‘Farm’, and music creation tool ‘MusicalM’. Bard is a light-weight version of lambda.

Consequently, predictions poured in that Google would mix bard with search just as Microsoft (MS) combined ‘Chat GPT’ with search engine ‘Bing’.

Google bard demonstration screen (Photo = Google)
Google bard demonstration screen (Photo = Google)

On the eighth of last month, at an illustration held in Paris, France, search manager Prabhakar Raghavan showed a search case using bard.

Nonetheless, on the event that day, the bard gave a fatally flawed answer and suffered a setback. Shares of Alphabet, Google’s parent company, fell 7%.

Since then, Google has been known to experiment with combining search and AI chatbots without disclosing test results. Reports followed, comparable to Sundar Pichai CED instructing all employees to factually correct Bard’s answers, and that Google was testing a search page design incorporating Bard.

On this process, Google executives appear to have judged that combining search and AI chatbots, like MS, is very prone to lead to a side effect of inaccuracy, which is the strength of Google search.

It’s a typical assessment amongst experts that AI chatbots are interactive, so they have a tendency to make up facts that will not be programmed to reply questions and provides plausible answers. Consequently of Microsoft’s direct integration of ChatGPT with Bing, cases by which chatbots submit incorrect answers are being reported one after one other.

CEO Sundar Pichai also said at a gathering on the 2nd, “It is crucial to acknowledge that Bard is experimental,” and “It’s much more essential to acknowledge its limitations.”

Jeong Byeong-il, member jbi@aitimes.com

LEAVE A REPLY

Please enter your comment!
Please enter your name here