How to Maximize the Positives of Generative AI and GPT

UCLA Anderson lecturer Arun Rao (’13), lead project manager in generative AI at Meta, joined Easton Technology Management Center faculty director Terry Kramer for a talk titled Generative AI and GPT: The Latest Developments Unveiled. The conversation, presented by the Easton Center, offered an update to Rao’s popular presentation last fall, in which he introduced generative AI as a transformational form of AI that can “create” new content and ideas, including conversations, stories, images, videos and music.

Here are some highlights of their conversation:

Accelerate Now

According to Rao, acceleration in generative AI is taking place in three-to-nine-month cycles. Previously, the doubling of AI probably took 18 months, but now there are subcycles that aren’t very long. Rao notes that if a company pauses a project for a year, it could find itself far behind in terms of the level of artificial intelligence it has to work with.

Scaling Laws Are a Big Deal

In AI, scaling laws mean that if you increase the size of a network and the amount of amount of data you’re feeding it, you get an increase in performance. Rao says that scaling laws are a big deal, and the trend is in cheap, mass-available intelligence through which “you could have an army of automated workers behind you to do anything you want.”

Rao says scaling up affects everyone, everywhere. It allows everyone to become more productive, from the individual to small businesses to large companies, and we need to be very thoughtful about it.

“The main thing is to use all the tools, use all the agents. Disrupt yourself. You’ll often find when you start using the tools, maybe once a month, then once a week, then daily, then multiple times a day, it just takes effort to get to understand them and know how they fit into your workflows. The economic value is great.”

Rao believes the impact on civilization is going to be much larger than we expect. “We all have the burden to figure out how to maximize the positives and also minimize the costs that come from AI in a thoughtful way.”

What Does It Mean for Consumers, Business and Society?

In his summary of the conversation, Kramer mentioned three categories of those who will be affected by generative AI technology: First, individuals, whether they are consumers or workers. Next, companies. And third, society.

For individuals, AI is going to change products and services as well as user experiences, whether through virtual assistants or increased knowledge management capabilities. “The impact on workers is going to be profound,” Kramer says. “AI is not going to take people’s jobs but those who use AI will take jobs away from people who don’t use AI.”

Companies that have been “digital first” are going to have to become “AI first.” Many companies are going to fall behind if they are not AI first, including some legacy companies. As a result, we are going to see changes in relative competitive positions.

The impact on society, according to Kramer, is actually very similar to the impact on individuals and companies. There will be changes in outcomes depending on how technology and AI are managed.

“Basically, the message is, whether it’s an individual, a company or society at large, you better get in and use this stuff,” says Kramer. “That doesn’t mean that we don’t need guardrails. It doesn’t mean you don’t have to be thoughtful about product design. But hanging back can actually create a much bigger negative outcome; broadly, a divide between workers, companies and societies who use or don’t use this technology.”