The Prompt: Generative AI industry trends

Business leaders are buzzing about generative AI. To help you keep up with this fast-moving, transformative topic, each week in “The Prompt,” we’ll bring you observations from our work with customers and partners, as well as the newest AI happenings at Google. In this edition, Philip Moyer, Global VP, AI & Business Solutions at Google Cloud, and Carrie Tharp, VP, Strategic Industries at Google Cloud, discuss industry trends from several recent trade shows. 

With most business leaders now acquainted with the broad potential of generative AI, executives we talk to are increasingly eager to understand the uses and opportunities specific to their industry or organization. 

That’s certainly been our experience over the last few weeks, as Google representatives met with scores of leaders at three of the biggest trade conferences: NAB Show for media, Hannover Messe for manufacturing and industrial development, and HIMSS for healthcare. Here are some key insights from these conversations.

Media leaders see AI not only generating content, but also reshaping relationships with audiences 

At NAB, some of the buzziest sessions focused on how generative AI can impact media production and management workflows. This topic has since stretched beyond the convention hall, with Joe Russo, co-director of the last two Avengers movies, predicting AI could revolutionize film production in as little as two years. 

AI-generated content and AI-enabled workflows are just part of the media discussion, however. We recently interviewed TIME senior vice president for data, product, and engineering Burhan Hamid about how generative AI could change the ways media companies build relationships and communities with consumers—check out the conversation here. 

We’ve also been talking to media leaders about leveraging generative AI for not only content production, but also enhanced personalized experiences and expanded content monetization opportunities—see this blog for the full scoop. 

Generative AI inspires excitement among healthcare and life sciences leaders—but security and regulatory concerns loom large

Generative AI’s potential in healthcare is obvious. For example, AlphaFold, DeepMind’s revolutionary system for predicting protein geometry, relies on the Transformer architecture, just like the generative AI models gaining prominence today in services like Bard. AlphaFold is opening new possibilities for drug discovery, and healthcare leaders are enthusiastic about the future. Our new Med-PaLM 2 model, the first large model to perform at an “expert” test-taker level performance on the MedQA dataset of U.S. Medical Licensing Examination (USMLE)-style questions, drove a lot of excitement and questions at HIMSS. 

Still, many executives remain cautious in this highly-regulated and data-sensitive space, and they expect cloud providers to work closely with industry leaders to make compliance simple and keep data safe. 

Manufacturers see opportunities from the factory floor to e-commerce 

At Hannover Messe, a steady stream of attendees came by our booth to talk about the art of the possible when it comes to bringing AI into the factory today, and just as many were curious about what generative AI portends for their industry. 

It’s still early days, so like many of the executives we met at HIMSS, leaders at Hannover Messe were concerned about regulatory issues and pleased to learn Google Cloud is prioritizing data sovereignty, auditing, access control, and other key capabilities as it rolls out enterprise-grade generative AI services. 

Machine-generated monitoring was one major talking point. Manufacturers have done a good job connecting their machines and collecting data, but making sense of it has been a far bigger challenge. AI can be a big help. Potentially, instead of scrolling through thousands of lines of measurements over weeks, puzzling out tolerances and other measurements of misalignment, a factory employee could simply ask an AI a question like, “What is the mechanical health of the line, and are there any parts we need to adjust or replace in the next 48 hours?”

Many attendees were also intrigued by generative AI use cases that are inspiring excitement more broadly, such as document search and summary assistants, improved e-commerce experiences, and more effective customer service. 

It’s not just about technology—execution and compliance matter

Earlier in this blog series, we mentioned that though generative AI is inspiring excitement in C-suites, many leaders have also expressed data and security concerns, with some organizations already restricting how or if employees can adopt certain services. Our trade show conversations about regulatory compliance, security, and reliability reinforce this point. 

Indeed, governments and regulators have started applying scrutiny too. Spain's data protection authority is investigating OpenAI for potential GDPR breaches, for example, and in the U.S., Senator Chuck Schumer is drafting a framework for AI guidelines to address national security. We’ll have more on this important topic in future editions of “The Prompt”—but suffice to say, it’s more important than ever for businesses to form AI principles to guide their strategy, and to ensure their technology providers have done the same. 

This week in AI at Google 

What happens when you combine a brain with a mind? Many Google teams have contributed to AI advancements over the years, but Google Brain and DeepMind have been widely recognized as some of the most influential, not just within Google but across the world. With the pace of AI innovating accelerating, we’re bringing these two teams together to form Google DeepMind—see this blog post to learn how this change helps us to boldly and responsibly deliver AI to our customers. 

A poet? A developer? Why not both? Since we launched Bard, users have been generating poems, blogs, answers to questions, and many other types of content—and with our latest update, they can now get assistance with coding too. As Google’s Paige Bailey explores in this blog post, Bard can help with programming and software development tasks across more than 20 languages, including C++, Go, Java, Javascript, Python, and Typescript.