The impact of artificial intelligence on work efficiency will depend on how we use this technology

Hi, I’m writing today from Deer Valley, Utah, where Fortune is organizing a conference called Brainstorm Tech. AI, unsurprisingly, was the main topic of the event. Here’s a summary of the most important AI tidbits so far:

On Monday, my colleague Emma Hinchliffe interviewed San Francisco Federal Reserve President Mary Daly, who said that the impact of generative AI on the job market will depend on what we do with the technology. Daly said we should expect generative AI to contribute at least to average productivity growth, which is currently 1.5% per year.

But she also said that if AI helps us invent new products and new processes, rather than simply automating existing ones, its potential impact on productivity growth will be much greater. “If we say that over the course of a decade (AI) has disappointed, that’s because of us,” Daly said. She also noted that all previous new technologies have created more jobs than they have eliminated over the long term, and she suspects AI will be no exception.

Taking up these ideas, Stanford University economist Erik Brynjolfsson has urged companies to view AI as a complement to human labor. While acknowledging that many companies struggle to find a way to get a reasonable return on investment from generative AI, Brynjolfsson said the key is to stop thinking about work and start thinking about tasks. AI can automate some tasks in an organization, but it can’t automate entire tasks (at least not yet). In fact, as automation helps reduce the costs associated with some roles, demand for those roles may actually increase, leading to more people being hired for those roles. (This is called the Jevons paradox.) Brynjolfsson is the co-founder of a company called Workhelix, which helps companies conduct this type of task-based analysis and develop strategic plans for implementing AI in the most impactful way possible across their organizations. Among the tasks best suited for AI automation are many in software development and customer service centers, he said.

Robinhood CEO Vladimir Tenev said Fortune Editor-in-Chief Alyson Shontell that she sees AI democratizing access to wealth management services. While ultra-high net worth individuals will continue to be served by human financial advisors, AI will be able to provide many others with access to good financial advice who would never have been able to afford a financial advisor before.

Agility Robotics CEO Peggy Johnson showed off the humanoid robot Digit, which is already working in warehouses under a multi-year agreement with GXO Logistics. Johnson said Agility is now integrating Digit with large language models (LLMs) so that humans can give Digit commands in natural language. Johnson said she believes Digit and humanoid robots like it are essential to help fill a shortage of about 1.1 million warehouse workers in the U.S.

Clara Shih, head of AI at Salesforce, spoke about how to build trust in AI in large organizations. She touted Salesforce’s Einstein AI Trust Layer, which includes features like data security and safeguards against generating toxic language, and defenses against prompt injection attacks, which is when an adversary creates a prompt that is designed to trick the LLM into bypassing its defenses.

She also said the company will soon begin rolling out AI software with more “agency” characteristics. These are AI models that can perform tasks within workflows, rather than just generate emails, letters or customer service dialogues. More broadly, Shih said one way organizations could build greater trust in AI is to make sure they’re using the right AI model for the problem at hand. Simply throwing a general big-language model at every business dilemma is unlikely to deliver the value companies expect from AI.

This morning I interviewed Google’s chief scientist, Jeff Dean, who said that increasingly longer context windows, like the ones Google promoted in Gemini, will help tame AI hallucinations. But he also agreed with recent comments from Microsoft’s Bill Gates that LLMs alone won’t deliver AGI, even if we continue to scale them. Dean agreed that some other algorithmic innovations will be necessary.

There will be a lot more AI discussion on Brainstorm Tech over the next few days, ending around noon Wednesday. You can join the live stream here , watch archived sessions here , and catch up on many of the sessions on fortune.com .

Here’s more news on AI.

Jeremy Kahn
[email protected]
@jeremyakahn

Before we get to the news…To better understand how AI can transform your business and hear from top business leaders in Asia on how AI is impacting different industries, join me on Fortune Brainstorm AI SingaporeThe event is taking place July 30-31 at the Ritz Carlton in Singapore. And today is your last chance to register! We have Satyen Sangani, CEO of Alation, who talks about the impact of AI on the digital transformation of Singapore’s GXS Bank, Sutten Thomas Pradatheth, CTO of Grab, who talks about how quickly AI can be implemented in the APAC region, Josephine Teo, Minister of Communications and Information of Singapore, who talks about the island nation’s drive to become an AI superpower, and much, much more. You can sign up here. Just for Eye on AI readers I have a special code that will allow you 50% discount about the registration fee. This is BAI50JeremyK.

This story was originally published on Fortune.com