[ad_1]
SAN FRANCISCO. Businesses and non-profit groups have come to a consensus after testing some of the latest advances in artificial intelligence: it is already changing the way they work.
The five organizations that were among the first to get access to GPT-4, the latest product from San Francisco-based startup OpenAI, said in interviews that they are reassigning employees, refocusing internal teams and rethinking their strategies in anticipation of technology transformation. most of their work. .
Their experience supports the idea that, for better or worse, AI technology could very soon radically change the daily lives of some people.
But organizations have also said the technology requires a huge amount of work to customize to suit their specific needs, with employees providing daily feedback on the software to teach it terminology and techniques specific to their fields, such as education or finance. OpenAI, best known for building the AI chatbot ChatGPT, can then integrate the data from that work into its own model to potentially improve its technology.
In essence, each of the early testers is a microcosm of what others might go through as access to GPT-4 expands.
“Right now the market is under the impression that you plug into these machines and they give you all the answers,” said Jeff McMillan, head of analytics, data and innovation, asset management at Morgan Stanley.
It’s not true, he said. He said the bank has 300 employees who spend part of their time testing their technology using GPT-4.
“We have a team of people literally looking at every response from the previous day,” he said.
For Morgan Stanley, the result is a custom chatbot built with GPT-4 that serves as an internal research tool for its financial advisors. McMillan said the tool has produced not only 60,000 research reports from various parts of the global economy, but also 40,000 other internal firm documents, making him an expert on any financial subject a financial advisor might want to find.
To be sure, the early adopters of the GPT-4 are not a random sample of the economy. OpenAI, which went commercial in 2019, has been choosing organizations in recent weeks and months.
Critics of OpenAI and its competitors argue that the AI sector has benefited from the non-skeptical hype over the past few months. According to founder Sal Khan, OpenAI was looking for positive examples when it approached Khan Academy, a non-profit educational organization, six months ago.
“The context was: we are going to work on a next generation model; we want to be able to launch it with positive use cases,” he said.
Khan Academy is best known for its YouTube videos, but since being approached by OpenAI, Khan said he has invested resources into building Khanmigo, a learning chatbot specifically trained in established learning concepts.
“Together we spent about 100 hours fine-tuning the model so that it could potentially behave like a really good tutor,” he said.
“If you look at the cost of tutoring, it can become a very, very big problem,” Khan added. “It’s like having a great graduate student, tutor or professor that you can start a conversation with at any time.”
Stripe, a tech company that makes payments software and related business products, said that when it got early access to GPT-4 in January, it fired 100 employees from their regular jobs and assigned them to an internal “hackathon” , in which each person spent an average week testing ideas.
Duolingo, a language learning app, got access to GPT-4 in the fall, and employees said that CEO Louis von Ahn was so engrossed in it that he called a meeting for 8 a.m. the next day and immediately changed people at work.
“After that, he said, ‘Change your team,’” said Edwin Bodge, product manager. “Since then, we have been working very closely with GPT-4 and with the OpenAI team.”
For now, Duolingo has added a new paid subscription tier for $29.99/month or $167.88/year that allows you to access a chatbot to speak French or Spanish. They’ve also added an AI bot that will explain grammar concepts to you as you progress through typical Duolingo lessons.
According to Boy, the company has created 1,000-2,000 verbal prompts for GPT-4 that make the bots work. The company will not share prompts upon request.
All of the organizations that spoke to NBC News said they acted with a degree of caution, given that AI technology is so new and the potential danger is unknown. Mike Buckley, CEO of Be My Eyes, a company developing an app for people who are blind or visually impaired, said he would like to have a test version of the GPT-4 app in more hands, “but we want to be thoughtful and safe.”
“Can we more broadly present this to the community in six to eight weeks? It’s possible, but we’re going to go where the data and use cases take us,” he said.
The company works by connecting visually impaired people with volunteers who, during a video call, can describe their surroundings to app users, such as a grocery store product label, airport directions, or the wording of a greeting card. . The GPT-4 version works without a volunteer on the other end because the AI describes what it “sees” with the camera.
According to a video she posted on TikTok, one of the app’s blind representatives used it to get directions on the London Underground system.
“We tried to hack it,” Buckley said, adding that his staff ran thousands of tests. “For several weeks we tried our best and were pleasantly surprised.”
He said his company didn’t run into security issues with GPT-4, but made mistakes; for example, confusing a toaster with a multicooker on the site.
[ad_2]
