AI is a fairly new technology, but its impacts are already being widely felt – and it will only become a bigger part of people’s personal and professional lives going forward. The Spring Eccles Alumni Forum helped paint a picture of the past, present, and future of AI with panelists Aaron Davis, Customer Engineer AI/ML at Google; Derek Egan, Product Manager for Vertex AI Google Cloud extensions; and Mat McBride, Corporate Vice President and CEO at Microsoft. The event was moderated by Justin Spangler, Enterprise Field Sales Representative – Strategic Growth Accounts at Google and a member of David Eccles Alumni Network Board.

Artificial intelligence, or AI, has evolved quickly, Davis said, starting with the advent of “big data.” For a long time, companies were occupied with collecting as much data as they could, he explained, but at some point, they had to figure out what to do with it. In other words, how can you derive value from data? That is the question AI is starting to answer.

And that progress has been accelerated, said McBride, by other technology and hardware breakthroughs such as GPUs (graphics processing units) that can compute large sets of data over a dispersed geographical area.

“We’re going to be sprinting trying to just consume this and find new ways to add new value from the technology,” McBride said. “It’s going to be fascinating to see.”

The key, he added, will be a holistic approach to AI that includes research, capabilities, and accessibility, such as apps that allow people to easily access AI solutions.

That ecosystem of tools is referred to as an agent, Egan said. Agents will further the mission of creating systems that can act autonomously, and AI will become applicable to most people through agents.

For example, Zebra Technologies, the scanning technology used in most retail stores, deployed an AI agent to index all the training documents and materials frontline retail workers are provided when they onboard to a new position.

All that information – for example, how to process a return – is available through the AI agent in every Zebra device. So, instead of an employee reading through hundreds of pages of processes and then trying to remember the relevant information when they start to process a return, the agent recognizes the task they are trying to complete and provides step-by-step instructions in real time.

And that’s just one example of what AI agents can do.

“That ecosystem as it comes together, I think, will really start to transform the way we interact with businesses and each other,” Egan said.

According to McBride, companies and individuals have barely scratched the surface of what AI can do. It’s exciting, but it also poses some risks. The technology is outpacing the ability of lawmakers to establish rules that ensure the models are fair, diverse, and unbiased so it’s up to companies, he said, to prioritize “responsible AI” and make sure their large data models are not surfacing skewed or incendiary data.

One solution is creating open-source models, Egan said. The more people who can contribute data to a model, the better chance it has of being representative.
People also shouldn’t trust everything that comes out of an AI model, Davis said, just like you wouldn’t trust everything that a co-worker or family member says. It’s the “trust but verify” principle. Users should make sure they can understand how the model got to the output, Davis said. Models can also be “grounded” with a smaller, validated set of data.

As for where AI is going, the panel agreed it is almost impossible to know. Models will continue to evolve and learn rapidly, they said, and use cases will expand as the number of sources for foundational data grows. Reasoning engines will also get better, they speculated, with smaller models targeted at specific use cases, like medical charting, for example, starting to proliferate.

One point of consensus? Wherever AI is going, it’s going there fast.

Want to learn more? Check out the full panel discussion at youtube.com/@EcclesAlumni/streams.