Microsoft executive believes AI chatbots

Need to learn how to express “help.”

 

 

 

According to Microsoft vice president Vik Singh, generative AI tools will save businesses a tonne of money and time even if the models have to learn to own up to when they don’t know what to do.

“Just to be really frank, the thing that’s really missing today is that a model doesn’t raise its hands and say ‘Hey, I’m not sure, I need help,'” Singh stated in an interview with AFP.

Since last year, ChatGPT and other generative AI applications have been rapidly deployed by Google, Microsoft, and its rivals. These apps generate a variety of content on demand and give users the impression that they are omniscient.

They still “hallucinate,” or make up responses, though.

The Copilot executive has a significant issue to resolve because Singh’s business clients cannot afford for even sporadic malfunctions in their AI systems.

Salesforce CEO Marc Benioff stated this week that he observed a growing number of his clients becoming dissatisfied with Microsoft‘s Copilot because to its ramblings.

According to Singh, “really smart people” are attempting to figure out how to make a chatbot acknowledge “when it doesn’t know the right answer and to ask for help.”

 

“Actual savings”

 

According to Singh, there would be no less value in a more modest model. “Tons of money” are saved even if the model has to consult a person in half of the cases.

At a client of Microsoft, “every time a new request comes in, they spend $8 to have a customer service rep answer it, so there are real savings to be had, and it’s also a better experience for the customer because they get a faster response.”

Singh began his employment with Microsoft in January. The AI assistant project “Copilot,” which focuses on online services, sales, and bookkeeping, has him leading the team as of July.

The onus is primarily on these apps to generate income and cover the significant generative AI expenses.

During the height of the artificial intelligence mania, businesses leading the field were bragging about extremely advanced systems that can “uplift humanity,” according to Sam Altman, the CEO of OpenAI, a company mostly funded by Microsoft.

Currently, though, the new technology’s main goal is to boost output and, preferably, profitability.

Microsoft claims that Copilot can conduct research on behalf of salespeople, freeing up time for customer calls. According to Singh, Lumen, a telecom business, “saves around $50 million a year” by doing this.

Singh’s teams are focusing on giving Copilot more autonomy and integrating it directly into the software of the massive tech company.

The executive said, “Let’s say I’m a sales representative and I have a customer call.” After two weeks, the model can “encourage the representative to follow up, or better yet, go ahead and send the email automatically on the representative’s behalf since it has been authorized to do so.”

“First inning” To put it another way, it’s expected that AI will relieve humans of tedious, repetitive tasks before it discovers a means to tackle global warming.

“We’re in the first inning,” Singh declared. “A lot of these things are productivity based, but they obviously have huge benefits.”

Will employment losses result from all of these productivity gains?

Prominent business executives, such as K Krithivasan of the massive Indian IT company TCS, have forecasted that contact centers will be nearly entirely replaced by generative AI.

However, Singh, like many other well-known figures from Silicon Valley, depends on technology to stimulate human creativity and perhaps provide new job prospects.

He related his 2008 Yahoo experience, in which twelve editors chose the stories that appeared on the home page.

“We came up with the idea of using AI to optimize this process, and some people asked ‘Oh my God, what’s going to happen to the employees?'” stated Singh.

Increased link clicks and the demand for new articles resulted from the automated system’s ability to renew content more quickly.

“Ultimately,” the CEO stated, “we needed to hire additional editors.”