Information & Ethics Committee on Dec. 13th, 2023

[René Villemure, Bloc, Trois-Rivières, QC]

Sorry to interrupt. I know you have a dozen policies. However, what can you tell us about them? What do we not know yet?

[Josh Harris]

When you're talking about foreign interference—I want to make sure that I'm as responsive as possible here—are you talking specifically about...? Can you give me a definition of foreign interference?

[René Villemure, Bloc, Trois-Rivières, QC]

Thank you, Mr. Harris, but I don't have much time left.

Ms. Hundley, I'm going to ask you the same question. You are responsible for influence operations policy. On the subject of foreign interference, could you tell us something we don't already know?

[Dr. Lindsay Hundley]

One thing I would highlight is that, as we noted in our latest transparency report earlier this year, we detected and removed a cluster of commenting activities from the influence operation known in the security community as “spamouflage” that targeted audiences in Canada. Researchers at the Australian Strategic Policy Institute have described that this operation's use of likely generative AI audio and doctored YouTube videos shared on other platforms “had zero or minimal engagement with real users.”

Spamouflage is a long-running, cross-Internet operation with global targeting that we and other industry peers have been enforcing on since 2019. In August, we removed thousands of accounts and pages under our CIB policies, after we were able to connect different clusters of activities together to be part of a single operation that we were able to attribute to individuals associated with Chinese law enforcement.

The one thing I will highlight here is that these operations are cross-Internet operations, so this particular activity, known as spamouflage, actually operated on over 50 platforms and forums across the Internet, including Facebook, Instagram, X or formerly Twitter, YouTube, TikTok, Reddit, Pinterest, Medium, Blogspot, LiveJournal, Vimeo and dozens of other smaller platforms, so we really believe that countering foreign interference is something that requires a whole-of-society effort.

[René Villemure, Bloc, Trois-Rivières, QC]

Thank you very much.

If you could provide that information in writing to the committee, it could be very helpful.

I'm going to put the same question to Mr. Huntley from Google.

[Shane Huntley]

Similar to what has already been spoken about, we've also been tracking “Spamouflage Dragon”, also known as “Dragonbridge”.

As has been pointed out, it's important to realize that many of these campaigns may be high in volume but very low in actual effects because of all the efforts to shut them down. It is actually much more difficult to get engagement on platforms as these foreign-coordinated actors than many people realize. I think part of that is due to the strong partnerships that we have, both with governments and across the industry here. This is not the first time many of us have met.

This work will be ongoing. We will apply it across the board and we will continually increase our understanding of this coordinated activity across our platforms.

[René Villemure, Bloc, Trois-Rivières, QC]

Ms. Hundley, could the development of artificial intelligence have an effect on foreign interference via social media?

[Dr. Lindsay Hundley]

That's an important question.

I think it's important to start by just noting what we have seen with regard to AI-generated content from foreign interference operations so far, and here I would note that the use of AI-generated content is not new. Since 2019, we have identified CIB operations on our platforms that have used profile pictures generated by a technique called generative adversarial networks, also known as GAN profiles. This use of AI-generated content doesn't allow these networks to evade our behaviour-based detections. In fact, over two-thirds of the CIB operations that we removed in the last year featured this type of AI-generated content.

We have seen newer operations using the latest generative AI techniques, and there are challenges that we anticipate there, including related to the scale of the content that can be created, but we fundamentally believe, from what we have seen, that a behaviour-based approach is still well suited for identifying covert influence operations early in their life cycle.

[The Chair, Conservative, John Brassard]

Thank you, Ms. Hundley.

[Dr. Lindsay Hundley]

This is because when we see these operations...by the time they post this content, they will have left a lot of behavioural signals that we can still detect, and I'm happy to provide more information on that.

[René Villemure, Bloc, Trois-Rivières, QC]

Thank you.

[The Chair, Conservative, John Brassard]

I'm sure we would love to hear more about that. Actually, I quite enjoy listening to you speak, Ms. Hundley. You have a lot of knowledge in this area. I can tell.

Mr. Green, you have five minutes. Please, go ahead.

[Matthew Green, NDP, Hamilton Centre, ON]

Thank you very much, Mr. Chair.

Mr. Chair, through you to Ms. Patell, in referencing Google, a lot of the conversation here today has been centred around the search platform and the public-facing interactions as they relate to the privacy of information.

I'm going to take a step back from that and reference the guiding principles on business and human rights that have been put forward by the Office of the United Nations High Commissioner for Human Rights.

Would you agree that companies have a responsibility to respect standards of international humanitarian law?

[Jeanette Patell]

Google is committed to respecting an approach to all of our products and services that puts user safety and privacy at the forefront.

[Matthew Green, NDP, Hamilton Centre, ON]

Do you have a responsibility to the standards of international humanitarian law?

[Jeanette Patell]

I'm not a lawyer, so I'm not in a position to be able to answer that.

[Matthew Green, NDP, Hamilton Centre, ON]

The international law would state that you do, and it would state that humanitarian law imposes obligations on business managers and staff not to breach the rules of international law.

I bring that up not in the context of your search engine platform but rather in the context of the cloud services that you provide. I want to hear from you on project Nimbus. As you may know, project Nimbus is a cloud computing project by the Israeli government and its military. It has come under condemnation for the use of cloud computing services, including artificial intelligence and machine learning, particularly as the project leads to furthering abuses of Palestinian human rights in the context of the ongoing occupation and the Israeli bombardment of Gaza.

Specifically to that point, I want to ask if you could just elaborate on Google's use of machine learning and artificial intelligence in what have been described by many UN experts as ongoing war crimes and crimes against humanity in Gaza.

[Jeanette Patell]

When it comes to project Nimbus, it's not a contract with which I'm personally familiar.

Google has said publicly that it is a project related to workloads that run on our commercial platform by Israeli government ministries such as finance, health care, transportation and education.

When it comes to—

[Matthew Green, NDP, Hamilton Centre, ON]

Would you agree that the primary contract in the Israeli government is the military and the use of artificial intelligence and machine learning as it relates to surveillance within the occupied Palestinian territories?

[Jeanette Patell]

No, I wouldn't agree with that. The work under project Nimbus is not directed at highly sensitive or classified military workloads that are relevant to weapons or intelligence services.

When it comes to our AI systems and products, all of our AI [Technical difficulty—Editor] products are anchored in our AI principles, which we were a leader in publishing in 2018, so we've been thinking about responsible AI development for a very long time.

One of the first principles that we have articulated is that AI needs to be developed for socially beneficial purposes designed to avoid reinforcing bias—

[Matthew Green, NDP, Hamilton Centre, ON]

Is it your testimony here today that your product is not being used for surveillance and unlawful data collection by the Israeli government? Is that your testimony here today?

[Jeanette Patell]

As I said, I can't speak to the specifics in that contract or how—

[Matthew Green, NDP, Hamilton Centre, ON]

Perhaps you could speak to the use of Google Ventures, the venture capital side of Google. Again, it's not a forward public-facing thing, but where the Alphabet company stashes its money. In particular, maybe you could speak to Project Maven and the use of drone technology there. Obviously, many of your former staffers have gone into these types of military contracts, which still have a connection with Google Ventures.

Can you just talk a little bit about Project Maven and its implications?

[Jeanette Patell]

Maybe I'll turn to my colleague, Shane Huntley, because again, this is not a project that I'm an expert on.

[The Chair, Conservative, John Brassard]

Shane, you have 40 seconds.