Google isn't a search company; it's an AI company.
That was the main message behind this year's Google I/O developer conference, which took place this past week at Shoreline Amphitheater in Mountain View, California. The I/O keynote was sprawling, covering everything from Gemini improvements to smart glasses. It's a lot to take in, but there are a few immediate takeaways that could change how nonprofit professionals work in the very near future.
Disclosure: TechSoup and Google have partnered on various initiatives over the years, and continue to do so. Additionally, Google has provided us with grant funding. This blog post was written and produced independently of those grants and initiatives and was not subject to Google's approval.
AI Is Everywhere
Throughout its I/O keynote, Google mapped out its strategy for integrating AI more deeply into everything it does — from Gmail to YouTube to Shopping to Search (more on that last one in a moment).
A lot of this iterates on the concept of the natural-language prompts that you're familiar with if you use a chatbot like Gemini, ChatGPT, or Copilot. For example, with the new Ask Maps feature, you can now ask Google Maps to "find me bike shops that specialize in e-bike repair in Oakland" and it will run a search for places that fit your criteria.
Another example that's more germane to nonprofits: I asked Maps to identify neighborhoods in my area that are two miles or more from a grocery store and it was able to highlight local food deserts on the map. While these use cases likely aren't what Google had in mind, these new features may provide nonprofits with some basic ways to better understand the current conditions in the communities they serve.
Like all generative AI–powered tools, these new features require a dose of caution. They may hallucinate or provide incorrect information. They also may not provide the same level of context that a human expert would, so it's best to use these sorts of features to test how effective they may be for you or as a starting point. In its keynote, Google itself seemed to frame many of its AI features — particularly around search — as tools to facilitate a deeper exploration of a topic rather than an end-all, be-all answers machine.
Making It Easier to Identify AI Content
Google announced that it is more fully adopting C2PA content credentials, an industry-standard method for identifying content that's been created or modified using AI. With the newest version of Gemini, you can quickly and easily verify whether an image or video clip is real or whether it's AI-generated or modified. Not all AI-generated content carries content credentials, but many of the largest AI firms are involved in C2PA's development.
Google also announced that OpenAI, Kakao, and ElevenLabs will adopt Google's own SynthID technology, which adds a digital watermark to AI content that is invisible to the human eye.
Although these announcements didn't get top billing during the keynote, they're significant steps forward in the fight against AI-generated misinformation and deepfakes.
Text-Based Prompts Are Just the Beginning
Right now, the most common form of interaction with AI is via chatbots. But Google made it clear that these text-based chat interfaces are only the beginning of what might be possible.
The clearest such example is Docs Live: It's a new feature that can take stream-of-consciousness spoken thoughts and shape them into a document that you can then iterate on, either manually or with further spoken instructions. If you find the task of crafting and then reworking written prompts challenging, you may find features like this easier to work with.
Agentic AI: Watch This Space
AI agents have received a lot of buzz over the last year or so. If you're unfamiliar with them, think of agents as mini-programs that can take actions on your behalf. Except, instead of coding them with a programming language, you create them using a prompt that defines what you want it to do. They're useful if you want to get notified about a change in a project's status, for instance, or to take care of busywork in the background.
Google is taking this concept and applying it to its services with a new app known as Gemini Spark. Google bills Spark as a personal AI agent that can process spoken and written prompts and process them all in the background, with little or no intervention on your part. One example from the keynote was an agent that pulled together information for a block party — a list of RSVPs, who's bringing what, and so on — and updated it on an ongoing basis.
Google isn't the first to offer this sort of thing — tools like OpenClaw have been around for a while — but it's another example of how AI is moving beyond chatbots and is working its way into more and more of our lives.
Google Search Fully Embraces AI — And Could Reshape the Web
Google announced a number of changes to its core web search, leaning hard into the capabilities of AI. There's a lot to talk about here. Let's get into it.
AI Mode
The first is AI Mode, a full-fledged chatbot-based search tool. It's live now, and you can access it by clicking the AI Mode button in the search bar on the Google home page. Enter a query or search prompt, and Google will provide you with answers and sources in a chatbot-style interface. You can also attach an image or file to go along with your query.
It's worth noting that traditional Google search is still there, and you'll get the more traditional search interface by default.
AI Mode search queries can be quite complex, and they can differ radically from a traditional keyword-based search. And the results can be quite sophisticated.
For example, a traditional keyword-based search would be something like, "California poverty line 2026." Meanwhile, AI Mode can support a search query like, "How has the poverty line changed in California since 2015? Show me with a line graph." AI mode will summarize the data and provide you with the graph that you requested, complete with source citations that you can use to verify the information presented.
From there, you can carry on the conversation and ask it to refine the search results or ask it an entirely new query.
The standard AI disclaimers apply: The information may be incorrect or incomplete. Use your own critical thinking skills and refer to primary sources to determine the veracity of any AI search results.
Agentic Search and Custom Search Dashboards
Agentic Search takes this a step further: With it, you'll be able to create agents to monitor pieces of information and act in a predefined manner based on it. In the keynote, Google's Liz Reid used the example of an agent that monitors the web for apartment listings that meet certain criteria and will send you a notification when it finds such listings.
Along those lines, Google is also bringing Antigravity, its AI-powered coding technology, to search, allowing you to create customized dashboards, interfaces, and visualizations for a wide range of information types.
These features arrive this summer for subscribers to Google AI Pro and AI Ultra plans.
Big Changes for the Web, but Uncertain Impacts for Nonprofits
Since it tends to keep users on Google's own platforms, the combination of a more robust AI mode and agentic search could accelerate a downward trend in search referral traffic to websites. This could disrupt the online economy as we know it.
As TechCrunch notes, "Combined, these changes will likely further decimate Google referrals to publishers, which have already been suffering from declining referrals due to AI Overviews. This has put some ad-dependent media operations out of business, and now things will likely get worse."
For nonprofits, the potential impact seems less clear. AI-enhanced search may make it easier for donors to find your organization, but then again, your organization may get buried beneath larger, more prominent nonprofits that work in your area of focus.
For example, an AI Mode search for "what are some good food banks to donate to in the San Francisco Bay Area?" surfaced several of the larger food banks in the region, but not the one located less than a mile away. Meanwhile, a traditional search provided a list of results that I could peruse on my own.
What Went Unsaid — And Where We Go from Here
What went unmentioned, though, was a discussion of energy consumption or commitments towards improving energy efficiency, nor did Google discuss the increase in datacenter expansion around the world necessary to fuel AI's progress.
A Reuters report from April stated that U.S. energy usage hit a record high in 2025 and that the Energy Information Administration expected further increases in 2026 and 2027. Depending on how we source that energy, AI's continuing growth could have significant impacts on efforts to address climate change.
We as a sector need to grapple with these sorts of concerns as we adopt AI — and as the technology becomes increasingly pervasive. If you'd like to learn more about mitigating climate impacts of AI, our Hive Mind project has an article that looks at what civil society organizations should consider as they adopt these emerging technologies.
[Thumbnail photo: Shutterstock]
