Prigozhin interests and Russian information operations

One of Threat Analysis Group’s (TAG) missions is to understand and disrupt coordinated information operations (IO) threat actors. Our research enables Google teams to make enforcement decisions backed by rigorous analysis. TAG’s investigations do not focus on making judgements about the content on Google platforms, but rather examining technical signals, heuristics, and behavioral patterns to make an assessment that activity is coordinated inauthentic behavior.

In this post, TAG is highlighting four case studies involving Russian IO tied to the Internet Research Agency (IRA) and its financier, Russian oligarch Yevgeny Prigozhin. In several cases, those campaigns served the dual purpose of promoting Russia’s agenda and Prigozhin’s business interests.

These examples underline broader trends we’re seeing: Russian IO groups are increasingly obscuring their role in influence operations, relying on stronger operational security and cutouts (intermediaries to mask their work) to dissociate themselves from user-facing activity. They launder their messages via local media brands, NGOs and PR firms that were in fact created by Russian shell companies. And in some cases, IRA-affiliated actors have responded to platforms’ enforcement efforts by moving to more permissive online spaces and platforms.

IO amplifying Prigozhin’s pro-Russian films

Prigozhin has financed several movies through a partial ownership stake in the film company, Aurum LLC. The company’s movies show Russia — especially the Russian military and mercenaries — in a positive light. The films have high production values and fictionalize Russia’s actions abroad in the style of Hollywood action movies. Storylines in the films include depictions of Russian soldiers in the Central African Republic, soldiers defending native Russians in Ukraine, and even a satire about the IRA and its role in the 2016 US elections. In 2021, they released “Солнцепёк” (“Sunlight” or “Blazing Sun” in English), which takes place in eastern Ukraine and claims to be a story based on true events from 2014 of Russian mercenaries, connected to the paramilitary Wagner Group, protecting Russians in Ukraine against Ukrainian forces.

Shortly after Russia’s invasion of Ukraine, TAG identified several IRA-affiliated news sites hosting ads to drive traffic to the videos including sites like newinform[.]com and slovodel[.]com. While the film was an older release from 2021, the timing of this campaign was notable because the subject matter mirrored newly topical real world events in Ukraine in a way that portrayed Russia positively. Google terminated nine new IRA-linked accounts using Ads to advertise the film and 44 new IRA-linked YouTube channels hosting clips, the full-length film and related comments. Some accounts claimed to be officially affiliated with the film, while others presented themselves as fan accounts.

A movie advertisement featuring the film's poster

Advertisement for the movie “Sunlight” on an IRA-affiliated news site

IRA-linked IO campaigns in Africa

In recent years, Russian IO actors tied to Prigozhin and the IRA, have peddled influence campaigns promoting the interests of Russia and Prigozhin’s Wagner Group in Africa. Researchers at Stanford, Graphika, and our colleagues at Meta have documented this trend going back to 2019. These campaigns involved creating NGOs, media brands and news agencies across Africa including a Ghanaian NGO, Sudan Daily, Peace Data and SADC News. These entities presented themselves as independent non-profit organizations and recruited local journalists and subject matter experts to publish content on topics like pro-Russia narratives, African pride and empowerment, and stories suggesting that Western imperialism is destroying Africa. Some authors likely did not realize they were working for a Russia-backed IO and genuinely believed in the content they wrote.

TAG’s investigations align with these earlier findings. Google terminated accounts and channels associated with the IRA’s fake media brands and NGOs throughout 2019 and 2020. This included IRA-linked accounts using Gmail to create profiles on non-Google social platforms, creating YouTube channels affiliated with the so-called news brands, and publishing content to Blogger.

In March 2021, Google shut down activity by several IRA-linked actors who published content promoting Wagner’s operations in Africa along with pro-Russia narratives. These articles appeared on Blogger and a number of non-Google blogging platforms such as Balalaika, Hashtap, Technowar and Voskhodinfo. The blogs amplified false narratives that the United Nations is funding terrorists in the Central African Republic and that Syrians need Wagner protection. The blogs were not backed by a social media presence.

a blog post showing soldiers in action

Example of a blog posted by an IRA-affiliated account

a blog heading showing a person holding a rocket launcher

Example of a blog posted by an IRA-affiliated account

In September 2022 Google terminated three IRA-linked YouTube channels that were sharing content in French and supportive of Russian policy objectives in Libya, including promoting a film in the Shugaley trilogy, another Aurum LLC film.

IRA influence operations concerning Ukraine

Russia’s agenda in Ukraine has also been a consistent, but not overwhelming, focal point for IRA-linked influence campaigns. In February 2022, Google terminated five YouTube channels and 21 Blogger blogs posting coordinated narratives on Blogger, YouTube and the Ukrainian blogging platform, Hashtap. In addition to domestically-focused content about Russia, several of the narratives focused on maligning Ukraine. These included allegations of Ukrainians deceiving Europe and stories of how Kyiv authorities failed to properly handle the Covid-19 pandemic. This activity spanned multiple blogging platforms and TAG observed the same IRA-linked accounts posted similar commentary across various news sites.

a muted and off-color flag is used at the top of a blog

IRA-created blog on Blogger criticizing EU support for Ukraine

IRA IO targeting domestic Russian audiences

Google regularly disrupts activity by IRA-linked accounts targeting Russian domestic audiences. These are often clusters of related accounts that create YouTube channels, upload videos, and comment and upvote each other’s videos. The activity occurs during Russian work hours, with narratives focused on Russian domestic issues and typically targeting political dissidents. In October 2022, Google terminated a cluster of nearly 700 IRA-linked accounts that were posting YouTube Shorts. The Shorts were crafted for a Russian domestic audience, praising Russian soldiers in Ukraine, and had negligible views or subscribers.

Other campaigns have focused on blogs. In July 2021, Google terminated 28 Blogger blogs created by IRA-linked accounts. Narratives in the blogs focused on Russian domestic affairs, including stories dismissing protests supporting anti-corruption activist, Alexei Navalny, denigrating local opposition politicians, criticizing the mayor of St. Petersburg and praising the heroics of Wagner Group. IRA actors also mirrored the same content on Ukrainian blogging platform, Hashtap. In some cases, multiple Blogger profiles published very similar or near-identical content.

The evolution of the Russian IO landscape

These case studies underscore several developments TAG observes in Russian IO activity. The accounts created lack well-developed, and backstopped personas, and increasingly are disrupted before they can gain traction. Russian IO actors also increasingly obscure their role, using stronger operational security and a range of intermediaries to conduct the actual user-facing activity. These proxies include third party PR firms, marketing agents, or unknowing local journalists and creators. Using well-selected proxies launders their legitimacy, and this provides an advantage compared to creating direct personas with little reach.

In our investigations of IRA-backed IO, we have also noted several cases where the narratives pushed by the IRA serve a dual purpose. Not only do they amplify messages supporting Russia, they also promote the business interests of oligarch, Yevgeny Prigozhin. Prigozhin has organized his empire around projects that directly and indirectly support the Russian state, and as the main financier of the IRA, he has cleverly leveraged his IO apparatus to amplify narratives that benefit not only Russia, but his own business interests as well.

Characterizing Emergent Phenomena in Large Language Models

Posted by Jason Wei and Yi Tay, Research Scientists, Google Research, Brain Team

The field of natural language processing (NLP) has been revolutionized by language models trained on large amounts of text data. Scaling up the size of language models often leads to improved performance and sample efficiency on a range of downstream NLP tasks. In many cases, the performance of a large language model can be predicted by extrapolating the performance trend of smaller models. For instance, the effect of scale on language model perplexity has been empirically shown to span more than seven orders of magnitude.

On the other hand, performance for certain other tasks does not improve in a predictable fashion. For example, the GPT-3 paper showed that the ability of language models to perform multi-digit addition has a flat scaling curve (approximately random performance) for models from 100M to 13B parameters, at which point the performance jumped substantially. Given the growing use of language models in NLP research and applications, it is important to better understand abilities such as these that can arise unexpectedly.

In “Emergent Abilities of Large Language Models,” recently published in the Transactions on Machine Learning Research (TMLR), we discuss the phenomena of emergent abilities, which we define as abilities that are not present in small models but are present in larger models. More specifically, we study emergence by analyzing the performance of language models as a function of language model scale, as measured by total floating point operations (FLOPs), or how much compute was used to train the language model. However, we also explore emergence as a function of other variables, such as dataset size or number of model parameters (see the paper for full details). Overall, we present dozens of examples of emergent abilities that result from scaling up language models. The existence of such emergent abilities raises the question of whether additional scaling could potentially further expand the range of capabilities of language models.


Emergent Prompted Tasks

First we discuss emergent abilities that may arise in prompted tasks. In such tasks, a pre-trained language model is given a prompt for a task framed as next word prediction, and it performs the task by completing the response. Without any further fine-tuning, language models can often perform tasks that were not seen during training.

Example of few-shot prompting on movie review sentiment classification. The model is given one example of a task (classifying a movie review as positive or negative) and then performs the task on an unseen example.

We call a prompted task emergent when it unpredictably surges from random performance to above-random at a specific scale threshold. Below we show three examples of prompted tasks with emergent performance: multi-step arithmetic, taking college-level exams, and identifying the intended meaning of a word. In each case, language models perform poorly with very little dependence on model size up to a threshold at which point their performance suddenly begins to excel.

The ability to perform multi-step arithmetic (left), succeed on college-level exams (middle), and identify the intended meaning of a word in context (right) all emerge only for models of sufficiently large scale. The models shown include LaMDA, GPT-3, Gopher, Chinchilla, and PaLM.

Performance on these tasks only becomes non-random for models of sufficient scale — for instance, above 1022 training FLOPs for the arithmetic and multi-task NLU tasks, and above 1024 training FLOPs for the word in context tasks. Note that although the scale at which emergence occurs can be different for different tasks and models, no model showed smooth improvement in behavior on any of these tasks. Dozens of other emergent prompted tasks are listed in our paper.


Emergent Prompting Strategies

The second class of emergent abilities encompasses prompting strategies that augment the capabilities of language models. Prompting strategies are broad paradigms for prompting that can be applied to a range of different tasks. They are considered emergent when they fail for small models and can only be used by a sufficiently-large model.

One example of an emergent prompting strategy is called “chain-of-thought prompting”, for which the model is prompted to generate a series of intermediate steps before giving the final answer. Chain-of-thought prompting enables language models to perform tasks requiring complex reasoning, such as a multi-step math word problem. Notably, models acquire the ability to do chain-of-thought reasoning without being explicitly trained to do so. An example of chain-of-thought prompting is shown in the figure below.

Chain of thought prompting enables sufficiently large models to solve multi-step reasoning problems.

The empirical results of chain-of-thought prompting are shown below. For smaller models, applying chain-of-thought prompting does not outperform standard prompting, for example, when applied to GSM8K, a challenging benchmark of math word problems. However, for large models (1024 FLOPs), chain-of-thought prompting substantially improves performance in our tests, reaching a 57% solve rate on GSM8K.

Chain-of-thought prompting is an emergent ability — it fails to improve performance for small language models, but substantially improves performance for large models. Here we illustrate the difference between standard and chain-of-thought prompting at different scales for two language models, LaMDA and PaLM.

Implications of Emergent Abilities

The existence of emergent abilities has a range of implications. For example, because emergent few-shot prompted abilities and strategies are not explicitly encoded in pre-training, researchers may not know the full scope of few-shot prompted abilities of current language models. Moreover, the emergence of new abilities as a function of model scale raises the question of whether further scaling will potentially endow even larger models with new emergent abilities.

Identifying emergent abilities in large language models is a first step in understanding such phenomena and their potential impact on future model capabilities. Why does scaling unlock emergent abilities? Because computational resources are expensive, can emergent abilities be unlocked via other methods without increased scaling (e.g., better model architectures or training techniques)? Will new real-world applications of language models become unlocked when certain abilities emerge? Analyzing and understanding the behaviors of language models, including emergent behaviors that arise from scaling, is an important research question as the field of NLP continues to grow.


Acknowledgements

It was an honor and privilege to work with Rishi Bommasani, Colin Raffel, Barret Zoph, Sebastian Borgeaud, Dani Yogatama, Maarten Bosma, Denny Zhou, Donald Metzler, Ed H. Chi, Tatsunori Hashimoto, Oriol Vinyals, Percy Liang, Jeff Dean, and William Fedus.

Source: Google AI Blog


Continuing our Commitment to User Choice Billing

Posted by Paul Feng, Vice President, Product ManagementBuilding on Android’s long history of continuously evolving to provide users and developers more flexibility and choice, we announced earlier this year that we would begin exploring expanded billing options on Google Play through our user choice billing pilot. At the heart of this pilot is our belief that the best way to offer alternative billing for in-app purchases is to put the choice in the hands of users.

Pilot participants can offer an additional billing system alongside Google Play’s billing system for their users in select countries. Our goal is to understand complexities involved in supporting user choice billing for developers and users in countries across the world  while maintaining a safe and positive user experience. This pilot allows us to test and iterate on different implementations, and gather insights from developers and users on their experience to determine how this pilot might evolve.
Illustration of a woman standing in front of a large phone with Google Play logo prominently featured. At her feet are stacks of coins. To the right of the phone are icons for music, video, a key, gems, and stars
Learn more about the user choice billing pilot here

Partner participation and excitement

When we announced the pilot, we noted that we were starting with Spotify as our very first partner. We’ve been working closely with the Spotify team and are excited to announce that this week they begin rolling out an initial test implementation of user choice billing to their users in select countries. We expect the experience will likely evolve over time as they continue to iterate and learn. Spotify has announced more detail on this rollout here.

We’re also excited that Bumble has joined to partner with us in our user choice billing pilot. We’re working with their teams and we anticipate their users will begin seeing this choice in-app in select countries in the coming months.

Enable user choice billing in over 35 countries

Additionally, with strong interest from developers around the world, in September we opened participation in the pilot to to all developers of non-gaming apps. We provided more detail about the eligibility, requirements—including interim UX guidelines—and announced that user choice billing will initially be available to users in Australia, India, Indonesia, Japan, and the European Economic Area.

Today we are excited to announce that based on the positive response and initial feedback from developers and users, we are expanding the pilot to users in the United States, Brazil, and South Africa.
Greyscaled world map highlighting existing pilot countries in green (Australia, India, Indonesia, Japan, European Economic Area) and recently added pilot countries in blue (Brazil, South Africa, United States)
Participating developers can enable user choice billing in over 35 countries

While this is still early days in the pilot, we’re encouraged by this initial response and momentum, and look forward to sharing more in the coming months as we continue to build and iterate with our partners and roll out user choice billing to more users. To learn more about the pilot eligibility, requirements, and how to get started visit our Help Center.

Feed label support added to datafeeds service in Content API for Shopping

On September 22, 2022, we updated you on changes to country targeting for shopping products, and how to use the feedLabel field. We’ve made additional changes to help you integrate feedLabel. Here are our previous announcements: What’s new
Merchant Center & Content API
As of November 8th, 2022 we’ve added the ability to manage feedLabel for datafeeds. The feedLabel field is now available in the following resources:
  • products
  • datafeeds
  • DatafeedStatus
You can now see which countries a datafeed explicitly targets in datafeedtarget. This applies when you use feedLabel instead of country in the datafeedtarget configuration.

We’ve also added the targetCountries field for datafeeds, so you can configure targeting for datafeeds directly. You can still configure targeting outside the feed, for example, by setting the shipping attribute of the products resource.

Note: You can’t manage Primary and Supplemental API feeds with the datafeeds service. You need to use the Merchant Center UI.

Behavior changes
Here’s a clarification of new API behavior for feedLabel:

Insert and update
You can now call Products.insert and Products.update with a feedLabel set to any valid string, for example “WINTERPRODUCTS”.

You can now use feedLabel without setting targetCountry during insertion and updates. Errors that used to warn of this requirement have been removed.

If you use both feedLabel and targetCountry in these calls, their values must be the same.

See Use feed labels to advertise products from specific feeds for the definition of a valid string for feedLabel.

Targeting
If you don’t use targetCountry for products, you must either set the shipping attribute of the products resource, or use the targetCountries field for the datafeeds resource to ensure your products target the chosen countries.

Opt out of receiving products and datafeeds without a country
If you’re concerned your codebase cannot handle products and datafeeds without a country, and you want to opt out of receiving them via the Content API for Shopping, fill out the following form: Feed label replaces target country in the Content API for Shopping - temporary exemption.

When you’re ready to support feedLabel, you can opt back in to receiving these offers.

If you have any questions about this change, please visit the Content API for Shopping forum.
Ian Ballantyne, Content API for Shopping Team

Feed label support added to datafeeds service in Content API for Shopping

On September 22, 2022, we updated you on changes to country targeting for shopping products, and how to use the feedLabel field. We’ve made additional changes to help you integrate feedLabel. Here are our previous announcements: What’s new
Merchant Center & Content API
As of November 8th, 2022 we’ve added the ability to manage feedLabel for datafeeds. The feedLabel field is now available in the following resources:
  • products
  • datafeeds
  • DatafeedStatus
You can now see which countries a datafeed explicitly targets in datafeedtarget. This applies when you use feedLabel instead of country in the datafeedtarget configuration.

We’ve also added the targetCountries field for datafeeds, so you can configure targeting for datafeeds directly. You can still configure targeting outside the feed, for example, by setting the shipping attribute of the products resource.

Note: You can’t manage Primary and Supplemental API feeds with the datafeeds service. You need to use the Merchant Center UI.

Behavior changes
Here’s a clarification of new API behavior for feedLabel:

Insert and update
You can now call Products.insert and Products.update with a feedLabel set to any valid string, for example “WINTERPRODUCTS”.

You can now use feedLabel without setting targetCountry during insertion and updates. Errors that used to warn of this requirement have been removed.

If you use both feedLabel and targetCountry in these calls, their values must be the same.

See Use feed labels to advertise products from specific feeds for the definition of a valid string for feedLabel.

Targeting
If you don’t use targetCountry for products, you must either set the shipping attribute of the products resource, or use the targetCountries field for the datafeeds resource to ensure your products target the chosen countries.

Opt out of receiving products and datafeeds without a country
If you’re concerned your codebase cannot handle products and datafeeds without a country, and you want to opt out of receiving them via the Content API for Shopping, fill out the following form: Feed label replaces target country in the Content API for Shopping - temporary exemption.

When you’re ready to support feedLabel, you can opt back in to receiving these offers.

If you have any questions about this change, please visit the Content API for Shopping forum.
Ian Ballantyne, Content API for Shopping Team

Saving water in L.A., one leaky toilet at a time

In water-scarce regions like California, every last drop counts. Yet millions of gallons of water are lost every year to a common, yet easily preventable, cause of water waste: leaky toilets.

That's why we recently co-funded a pilot project to install water-saving technology in three multi-family buildings in Los Angeles. The tech takes aim at common leaks, like toilets that keep running water when not in use, which can add up over time. The pilot is on track to save 6.4 million gallons of water a year in the L.A. watershed where we operate, supporting our commitment to replenish 120% of the water we consume, on average, across our offices and data centers by 2030.

The pilot came together with partners from the California Water Action Collaborative (CWAC), a water stewardship network of over 25 organizations — including private companies like Google alongside environmental NGOs and nonprofits — that are committed to improving water security across the state.

Here's a look at how this project is saving water, money and energy, and at the potential for collective action models to make meaningful progress on rising water challenges.

Saving water, money and energy

The Los Angeles Department of Water and Power estimates that the average household loses up to 10,000 gallons of water every year to leaky toilets that go unnoticed. The good news is that while leaky toilets can be hard to detect, they’re easy to fix.

For the pilot project, CWAC members Pacific Institute and Bonneville Environmental Foundation tackled this challenge in three low-income multi-family housing buildings operated by nonprofit organizations, working alongside the Metropolitan Water District of Southern California and other local water utilities. Toilets in these buildings were equipped with small, low-cost, low-power sensors developed by Sensor Industries. When a toilet leaks, the sensors alert building management in real time that a toilet needs to be repaired. The fix is usually as simple as readjusting or replacing the toilet flapper.

This simple intervention resulted in serious savings of water, money and energy, according to estimates from the nonprofit Pacific Institute:

  • Water: The pilots are reducing building water use by an estimated 15% to 25%. The expected savings of 6.4 million gallons of water per year is equivalent to the total annual water use of about 40 single-family homes. Those savings extend to other customers who get their water from the same public utility, reducing water demand — and improving water reliability and affordability — across the system.
  • Cost: The water savings translate into cost savings on water and wastewater bills of the same 15% to 25%, amounting to tens of thousands of dollars a year. The nonprofit building operators who pay the water bills could use these savings to make building improvements, in effect passing the savings along to residents.
  • Energy: Southern California imports much of its water from hundreds of miles away, and it takes a lot of energy to pump this water over the mountains surrounding the L.A. Basin and treat it for household use. By reducing the demand for that water, the project cuts back on the energy and associated greenhouse gas emissions embedded in the water system.

Pacific Institute points to several other advantages of this approach. Residents don’t have to do anything — the non-invasive system detects problems and notifies the building. Facility managers can see the likely reason for the leak (such as a stuck flapper), which helps them fix it faster. The nonprofit building operators can focus on more urgent issues and reduce time spent tracking down leaks.

Bringing the solution to more cities

Taking this pilot to other places has always been a goal, and that expansion effort is now underway. We’re funding work to bring this solution to a 225-unit building in San Francisco that shares a watershed with our local offices. Here we expect to save a little over 1 million gallons of water a year, based on the savings found in L.A.

In New York City, we’re exploring this approach in a building a few miles from our main local campus, and here too we expect to save roughly 1 million gallons of water a year. While this region is not currently in a drought, we expect the system to save significant amounts of energy, as New York City imports its water from far away. Additionally, this project can help reduce pressure on New York’s combined waste- and stormwater system, which can overflow into clean waterways during heavy storms.

In the face of difficult decisions around water resources and scarcity, it’s not easy to find meaningful wins that everyone can get behind. The pilots represent a solution that local utilities anywhere can adopt with the right partners.

Looking ahead, we’ll continue to support collective action around watershed health in the communities where we operate. A healthy, resilient water system takes all of us.

Related Article

Our commitment to water stewardship

We’re committing to replenish 120% of the water we consume, on average, by 2030 and supporting water security in areas where we operate.

Read Article

Saving water in L.A., one leaky toilet at a time

In water-scarce regions like California, every last drop counts. Yet millions of gallons of water are lost every year to a common, yet easily preventable, cause of water waste: leaky toilets.

That's why we recently co-funded a pilot project to install water-saving technology in three multi-family buildings in Los Angeles. The tech takes aim at common leaks, like toilets that keep running water when not in use, which can add up over time. The pilot is on track to save 6.4 million gallons of water a year in the L.A. watershed where we operate, supporting our commitment to replenish 120% of the water we consume, on average, across our offices and data centers by 2030.

The pilot came together with partners from the California Water Action Collaborative (CWAC), a water stewardship network of over 25 organizations — including private companies like Google alongside environmental NGOs and nonprofits — that are committed to improving water security across the state.

Here's a look at how this project is saving water, money and energy, and at the potential for collective action models to make meaningful progress on rising water challenges.

Saving water, money and energy

The Los Angeles Department of Water and Power estimates that the average household loses up to 10,000 gallons of water every year to leaky toilets that go unnoticed. The good news is that while leaky toilets can be hard to detect, they’re easy to fix.

For the pilot project, CWAC members Pacific Institute and Bonneville Environmental Foundation tackled this challenge in three low-income multi-family housing buildings operated by nonprofit organizations, working alongside the Metropolitan Water District of Southern California and other local water utilities. Toilets in these buildings were equipped with small, low-cost, low-power sensors developed by Sensor Industries. When a toilet leaks, the sensors alert building management in real time that a toilet needs to be repaired. The fix is usually as simple as readjusting or replacing the toilet flapper.

This simple intervention resulted in serious savings of water, money and energy, according to estimates from the nonprofit Pacific Institute:

  • Water: The pilots are reducing building water use by an estimated 15% to 25%. The expected savings of 6.4 million gallons of water per year is equivalent to the total annual water use of about 40 single-family homes. Those savings extend to other customers who get their water from the same public utility, reducing water demand — and improving water reliability and affordability — across the system.
  • Cost: The water savings translate into cost savings on water and wastewater bills of the same 15% to 25%, amounting to tens of thousands of dollars a year. The nonprofit building operators who pay the water bills could use these savings to make building improvements, in effect passing the savings along to residents.
  • Energy: Southern California imports much of its water from hundreds of miles away, and it takes a lot of energy to pump this water over the mountains surrounding the L.A. Basin and treat it for household use. By reducing the demand for that water, the project cuts back on the energy and associated greenhouse gas emissions embedded in the water system.

Pacific Institute points to several other advantages of this approach. Residents don’t have to do anything — the non-invasive system detects problems and notifies the building. Facility managers can see the likely reason for the leak (such as a stuck flapper), which helps them fix it faster. The nonprofit building operators can focus on more urgent issues and reduce time spent tracking down leaks.

Bringing the solution to more cities

Taking this pilot to other places has always been a goal, and that expansion effort is now underway. We’re funding work to bring this solution to a 225-unit building in San Francisco that shares a watershed with our local offices. Here we expect to save a little over 1 million gallons of water a year, based on the savings found in L.A.

In New York City, we’re exploring this approach in a building a few miles from our main local campus, and here too we expect to save roughly 1 million gallons of water a year. While this region is not currently in a drought, we expect the system to save significant amounts of energy, as New York City imports its water from far away. Additionally, this project can help reduce pressure on New York’s combined waste- and stormwater system, which can overflow into clean waterways during heavy storms.

In the face of difficult decisions around water resources and scarcity, it’s not easy to find meaningful wins that everyone can get behind. The pilots represent a solution that local utilities anywhere can adopt with the right partners.

Looking ahead, we’ll continue to support collective action around watershed health in the communities where we operate. A healthy, resilient water system takes all of us.

Related Article

Our commitment to water stewardship

We’re committing to replenish 120% of the water we consume, on average, by 2030 and supporting water security in areas where we operate.

Read Article

Get ready for Google Summer of Code 2023!

We are thrilled to announce the 2023 Google Summer of Code (GSoC) program and share the timeline with you to get involved! 2023 will be our 19th consecutive year of hosting GSoC and we could not be more excited to welcome more organizations, mentors, and new contributors into the program.

With just three weeks left in the 2022 program, we had an exciting year with 958 GSoC contributors completing their projects with 198 open source organizations.

Our 2022 contributors and mentors have given us extensive feedback and we are keeping the big changes we made this year, with one adjustment around eligibility described below.
  • Increased flexibility in project lengths (10-22 weeks, not a set 12 weeks for everyone) allowed many people to be able to participate and to not feel rushed as they wrapped up their projects. We have 109 GSoC contributors wrapping up their projects over the next three weeks.
  • Choice of project time commitment there are now two options, medium at ~175 hours or large at ~350 hours, with 47% and 53% GSoC contributors, respectively.
  • Our most talked about change was GSoC being open to contributors new to open source software development (and not just to students anymore). For 2023, we are expanding the program to be open to students and to beginners in open source software development.
We are excited to launch the 2023 GSoC program and to continue to help grow the open source community. GSoC’s mission of bringing new contributors into open source communities is centered around mentorship and collaboration. We are so grateful for all the folks that continue to contribute, mentor, and get involved in open source communities year after year.

Interested in applying to the Google Summer of Code Program?

Open Source Organizations
Check out our website to learn what it means to be a participating organization. Watch our new GSoC Org Highlight videos and get inspired about projects that contributors have worked on in the past.

Think you have what it takes to participate as a mentor organization? Take a look through our mentor guide to learn about what it means to be part of Google Summer of Code, how to prepare your community, gather excited mentors, create achievable project ideas, and tips for applying. We welcome all types of open source organizations and encourage you to apply—it is especially exciting for us to welcome new orgs into the program and we hope you are inspired to get involved with our growing community.

Want to be a GSoC Contributor?
Are you new to open source development or a student? Are you eager to gain experience on real-world software development projects that will be used by thousands or millions of people? It is never too early to start thinking about what kind of open source organization you’d like to learn more about and how the application process works!

Watch our new ‘Introduction to GSoC’ video to see a quick overview of the program. Read through our contributor guide for important tips from past participants on preparing your proposal, what to think about if you wish to apply for the program, and everything you wanted to know about the program. We also hope you’re inspired by checking out the nearly 200 organizations that participated in 2022 and the 1,000+ projects that have been completed so far!

We encourage you to explore our website for other resources and continue to check for more information about the 2023 program.

You are welcome and encouraged to share information about the 2023 GSoC program with your friends, family, colleagues, and anyone you think may be interested in joining our community. We are excited to welcome many more contributors and mentoring organizations in the new year!

By Stephanie Taylor, Program Manager, and Perry Burnham, Associate Program Manager for the Google Open Source Programs Office

Optimized targeting launch in Display & Video 360 postponed

The launch of optimized targeting and deprecation of targeting expansion for display, video, and audio line items in Display & Video 360 have been postponed. Optimized targeting was previously announced to gradually launch for all Display & Video 360 partners from November 7 to November 9, 2022.

The changes in Display & Video 360 API behavior that were previously announced have also been postponed. The targetingExpansion field in the LineItem resource will continue to represent the targeting expansion feature.

We will announce a new date for these changes at a later date.
- Trevor Mulchay, Display & Video 360 API Team