Enterprise

Google Cloud’s top AI executive is caught between academic ideals and corporate reality

Talking with Andrew Moore, Google Cloud’s VP and GM for AI, reveals a disconnect between Google’s bold plan for a government AI data cloud and an academic’s goals for collaborative, global AI research.

Andrew Moore

One of Andrew Moore’s most important and visible tasks involves defining what a national AI research cloud might look like.

Photo: CMU

It’s nothing new to hear Google executives talk about happenings inside different divisions of the company as though they’re separate entities. But when the head of Google Cloud AI talks about the company’s detailed proposal for a government AI project that he is closely involved in as though it were conjured up in some distant universe, it’s confusing, to say the least.

In this case, that executive is Andrew Moore, Google Cloud’s vice president and general manager for AI and Industry Solutions. An academic at heart, Moore is reluctant to speak as a Googler when it comes to the company’s plan for a federally funded AI research cloud. But his arms' length separation from the company’s proposal reveals that Google views it just as much a political gambit as a cloud-services sales call to the government.

Moore was a computer science professor at world-renowned AI research university Carnegie Mellon before first joining Google in 2006 to head up its Pittsburgh engineering office. A fish out of water, perhaps, he returned in 2014 to the university as the dean of its School of Computer Science.

Now, Moore is back at Google; he still works from his digs in Pittsburgh, the city he calls “the center of the world for the most advanced forms of robotics and mathematical AI.” And although Moore was born in England and has a Ph.D. from the University of Cambridge, some say his accent has morphed over the years into something that sounds a bit less British and a bit more mid-Atlantic.

The ecosystem Moore is immersed in now that he’s back corporate-side — Moore has had no affiliation with Carnegie Mellon since he rejoined Google in 2019 — is one very familiar to an AI researcher: the cloud. At Google, Moore helps customers coax their machine-learning prototypes into active algorithmic systems operating in the wild. Those projects, like the academic research that gave birth to today’s productized cloud-based AI tools, require massive amounts of computing power and data, and that’s what the cloud is all about.

Right now, one of Moore’s most important and visible tasks involves defining what a possible national AI research cloud might look like. That involves deciding who will provide cloud services to AI researchers in the U.S., and how.

Google has particularly bold ideas about how this would-be National AI Research Resource might be built and how the company should be involved. The thing is, even though Moore is helping determine what the resource could be, he said he had nothing to do with the company’s proposal for it.

“My role is as a commissioner on this panel ... The panel did a request for information from a whole bunch of companies. And Google was one of the companies that responded, but I was not the author of that,” he said regarding the proposal for the controversial project. “My role within the commission was watching what many stakeholders — not just the cloud companies — but also other smaller and [medium-sized] companies are saying."

A Google exec at arms' length from Google

There’s little question that the modern world of data and AI driven by cloud computing requires a new approach to advanced research.

“There is an old-fashioned way of thinking of doing these very data intensive bits of research, which is, you put some servers in your laboratory, your lab at the university, and you download a whole bunch of data. And you do your work there, and you've got your grad students and yourself sitting there warmed by your servers, sitting next to you humming away,” he said.

That on-campus server scenario has become more and more rare, of course. “So as a dean, I would have new faculty members coming to me. When they're talking about their startup packages, they would talk about, 'Well, I need such and such a number of servers with this much disk space to operate.' But as time has gone by, that's turned into folks saying, ‘I need cloud credits.’”

A federal AI cloud and data repository could be a game-changer for academic researchers, but Google seems to recognize the political implications, too. Indeed, according to Moore, it was the company’s public policy team, its government affairs division, that put together Google’s NAIRR proposal. A Google Cloud spokesperson clarified that the document was drafted by “many teams at Google that care about the success of the task force,” but did not provide the names of those teams. Google Cloud does have a team dedicated to managing work for public-sector clients, for example.

Moore is Google Cloud’s top AI executive, but he distanced himself from that role when answering Protocol’s questions about Google’s approach to the NAIRR initiative. For instance, rather than explain why Google wants to partner in the project with AWS, Microsoft and other cloud providers, companies it competes with in the cutthroat cloud industry, he said, “I'm not going to talk about Google specifically, like it's particularly different from the others.”

Instead, the on-again, off-again university educator emphasized the lack of computing power and data resources available to academic AI researchers, and lamented the lure of corporate work. “Many of those academics, they look around, and they see opportunities for themselves in industry, doing perhaps closed research, where they would be able to get more done, but it's less useful for the country as a whole. So it's incredibly important — and this is something where you will see agreement throughout the commercial world in the United States — it's incredibly important that we support academic researchers,” he said.

Legitimate concerns and momentous decisions

Although Moore does not consider himself a Google mouthpiece in his position as a NAIRR task force member, he and the company have highlighted some of the very same data-security, quality and access issues.

“This is a very tough thing,” said Moore during an October NAIRR meeting. He was talking about the risk of data exfiltration, when an authorized person extracts data from a secured system and shares it with unauthorized third parties or moves it to an insecure place.

“I think this committee has a really big decision,” he continued, discussing the types of data and level of data granularity that could be available in the national research hub.

“If we’re going to do controlled data, which does not allow exfiltration, that means a lot of work by us, like tens of millions of dollars possibly of software engineering or contracts to someone to set that up. I’m on the fence between them. I could say, 'Let’s just go with safe old geology and weather and outdoor natural scenes with faces blurred and keep our life simple.' Or I could say, 'No, obviously if you’re going to work on diseases or something like that, you need patient data protected against exfiltration.' We’ve got a big, and, I think, momentous decision here, because I don’t think this exfiltration limitation thing is a small side issue for us.”

Google’s public policy team had something similar on its mind in its proposal to the task force. “Before making Google datasets publicly available for the open-source community, we spend hundreds of hours standardizing data and validating quality,” the company wrote in its response to a request for information by the task force. “This expensive error prone process, which is repeated for each analysis, not only becomes a barrier to the use of data, but also leads to problems of reproducibility in research questions,” it said, adding that “the success of a research initiative potentially involving sensitive data depends upon the ability to reliably credential users and provide granular access management.”

No worries, Google seemed to imply in its submission. Not only would the company take on the arduous task of preparing raw data flowing into the research cloud, but it would do it for free. Some have questioned Google’s motivations, arguing that getting first dibs on the raw data could grant the company privileged access to valuable information others wouldn’t see.

In his interview with Protocol, Moore said Google’s proposal presents “a very legitimate concern that there's an attempt to sort of control data like this.” However, he said, “There would be a bit more concern if we were in some world without any data rights, or discussion of data security: [if] a commercial company was licensed to take all this data, do what it wants with it and then put up its own interpretation of that data publicly.”

On the benefits of working with China

The NAIRR task force was established on recommendation from the National Security Commission on AI, which warned that without a full-fledged national effort to advance AI research in the U.S., the country could lose its leadership position in AI to China, becoming more vulnerable to AI-enabled threats. However, Moore downplayed the notion of an “AI race” between the U.S. and China.

“There's many notions of winning AI. Just as with the aerospace industry, or the cybersecurity industry, or quantum, or in the early days of engine manufacture, you're going to see international competition,” he said. “And it absolutely is the case that the U.S. funding agencies are inspired by helping keep the United States at the forefront in most areas of science and technology.”

Being at the forefront of science is one thing, but some— including people in military, cyber and data security, intellectual property law or civil and human rights arenas — believe the risk of China dominating AI advancements poses a grave threat that ought to limit collaboration in AI research and business activity between the two countries.

Moore, a scientist at his core — his Twitter bio reads, “I love Algorithms” — doesn’t see it that way.

“Among academia, for example, you will see plenty of cases where there are mutually supportive, friendly, creative bits of joint work going on between different countries. And so, I want to be clear that this national security commission was not indicating that there should be no joint AI research between continents,” he said. “I think it's actually considered to be a strong benefit and a chance to sort of bring people together if there is collaboration between different researchers in this area.”

Policy

Google is wooing a coalition of civil rights allies. It’s working.

The tech giant is adept at winning friends even when it’s not trying to immediately influence people.

A map display of Washington lines the floor next to the elevators at the Google office in Washington, D.C.

Photo: Andrew Harrer/Bloomberg via Getty Images

As Google has faced intensifying pressure from policymakers in recent years, it’s founded trade associations, hired a roster of former top government officials and sometimes spent more than $20 million annually on federal lobbying.

But the company has also become famous in Washington for nurturing less clearly mercenary ties. It has long funded the work of laissez-faire economists who now defend it against antitrust charges, for instance. It’s making inroads with traditional business associations that once pummeled it on policy, and also supports think tanks and advocacy groups.

Keep Reading Show less
Ben Brody

Ben Brody (@ BenBrodyDC) is a senior reporter at Protocol focusing on how Congress, courts and agencies affect the online world we live in. He formerly covered tech policy and lobbying (including antitrust, Section 230 and privacy) at Bloomberg News, where he previously reported on the influence industry, government ethics and the 2016 presidential election. Before that, Ben covered business news at CNNMoney and AdAge, and all manner of stories in and around New York. He still loves appearing on the New York news radio he grew up with.

Sustainability. It can be a charged word in the context of blockchain and crypto – whether from outsiders with a limited view of the technology or from insiders using it for competitive advantage. But as a CEO in the industry, I don’t think either of those approaches helps us move forward. We should all be able to agree that using less energy to get a task done is a good thing and that there is room for improvement in the amount of energy that is consumed to power different blockchain technologies.

So, what if we put the enormous industry talent and minds that have created and developed blockchain to the task of building in a more energy-efficient manner? Can we not just solve the issues but also set the standard for other industries to develop technology in a future-proof way?

Keep Reading Show less
Denelle Dixon, CEO of SDF

Denelle Dixon is CEO and Executive Director of the Stellar Development Foundation, a non-profit using blockchain to unlock economic potential by making money more fluid, markets more open, and people more empowered. Previously, Dixon served as COO of Mozilla. Leading the business, revenue and policy teams, she fought for Net Neutrality and consumer privacy protections and was responsible for commercial partnerships. Denelle also served as general counsel and legal advisor in private equity and technology.

Workplace

Everything you need to know about tech layoffs and hiring slowdowns

Will tech companies and startups continue to have layoffs?

It’s not just early-stage startups that are feeling the burn.

Photo: Kirsty O'Connor/PA Images via Getty Images

What goes up must come down.

High-flying startups with record valuations, huge hiring goals and ambitious expansion plans are now announcing hiring slowdowns, freezes and in some cases widespread layoffs. It’s the dot-com bust all over again — this time, without the cute sock puppet and in the midst of a global pandemic we just can’t seem to shake.

Keep Reading Show less
Nat Rubio-Licht

Nat Rubio-Licht is a Los Angeles-based news writer at Protocol. They graduated from Syracuse University with a degree in newspaper and online journalism in May 2020. Prior to joining the team, they worked at the Los Angeles Business Journal as a technology and aerospace reporter.

Entertainment

Sink into ‘Love, Death & Robots’ and more weekend recs

Don’t know what to do this weekend? We’ve got you covered.

Our favorite picks for your weekend pleasure.

Image: A24; 11 bit studios; Getty Images

We could all use a bit of a break. This weekend we’re diving into Netflix’s beautifully animated sci-fi “Love, Death & Robots,” losing ourselves in surreal “Men” and loving Zelda-like Moonlighter.

Keep Reading Show less
Nick Statt

Nick Statt is Protocol's video game reporter. Prior to joining Protocol, he was news editor at The Verge covering the gaming industry, mobile apps and antitrust out of San Francisco, in addition to managing coverage of Silicon Valley tech giants and startups. He now resides in Rochester, New York, home of the garbage plate and, completely coincidentally, the World Video Game Hall of Fame. He can be reached at nstatt@protocol.com.

Workplace

This machine would like to interview you for a job

Companies are embracing automated video interviews to filter through floods of job applicants. But interviews with a computer screen raise big ethical questions and might scare off candidates.

Although automated interview companies claim to reduce bias in hiring, the researchers and advocates who study AI bias are these companies’ most frequent critics.

Photo: Johner Images via Getty Images

Applying for a job these days is starting to feel a lot like online dating. Job-seekers send their resume into portal after portal and a silent abyss waits on the other side.

That abyss is silent for a reason and it has little to do with the still-tight job market or the quality of your particular resume. On the other side of the portal, hiring managers watch the hundreds and even thousands of resumes pile up. It’s an infinite mountain of digital profiles, most of them from people completely unqualified. Going through them all would be a virtually fruitless task.

Keep Reading Show less
Anna Kramer

Anna Kramer is a reporter at Protocol (Twitter: @ anna_c_kramer, email: akramer@protocol.com), where she writes about labor and workplace issues. Prior to joining the team, she covered tech and small business for the San Francisco Chronicle and privacy for Bloomberg Law. She is a recent graduate of Brown University, where she studied International Relations and Arabic and wrote her senior thesis about surveillance tools and technological development in the Middle East.

Latest Stories
Bulletins