At a briefing forward of Google I/O 2026, I watched firm execs unveil a listing of AI-powered options aimed toward fixing ache factors throughout its software program ecosystem.
One device guarantees to radically enhance the standard of video searches: Ask YouTube, because it’s known as, scours the platform’s catalog of long-form movies and Shorts to floor content material related to extra advanced queries.
At first look, that seems like a win for each YouTube viewers and creators. Ask YouTube, nonetheless, takes an additional step — it directs searchers to a video that solutions their question and zeroes in on the related timestamp. Get in, get your reply, get out.
But when folks do not stay to observe a complete video, or at the least most of it, this threatens the methods video creators earn money and construct a following. YouTubers want a big subscriber base, which ends up in extra advert income, sponsorships, affiliate hyperlinks and fan funding.
In different phrases, Google’s pursuit of person comfort might immediately defund the creators who energy its platform. Goodbye, advert income. So lengthy, neighborhood constructing.
How Google’s Ask YouTube function will look.
Google made an identical transfer when it launched AI Overviews in late 2024, rating readers over publishers that had lengthy relied on visitors from the search engine outcomes web page. By scraping info from media websites and summarizing it on the prime of search outcomes, AI Overviews scale back clicks to different web sites by 58%, in line with a February report by advertising and analysis agency Ahrefs. (Google claimed that “hyperlinks included in AI Overviews get extra clicks than if the web page had appeared as a standard internet itemizing for that question.”)
Within the opening minutes of final yr’s Google I/O 2025 keynote, Google CEO Sundar Pichai bragged that AI Overviews — a function that may’t be turned off and mechanically seems — had over 1.5 billion customers per 30 days. Alongside in the present day’s Google I/O 2026 keynote, Pichai penned a blog post noting that the function had grown to over 2.5 billion customers per 30 days.
The cannibalization of referral visitors by Google’s generative AI Overviews seems to be a predictive mannequin for video platforms. The Ask YouTube function, as described, will direct you to the purpose in a video answering their query, after which you will possible have little or no incentive to remain. Viewers possible will not get a way of the channel’s topic and vibe, nor will they preserve watching to interact with its attraction or storytelling.
Google’s “answer” is to surgically excise creator content material to funnel YouTube viewers to solutions. In essence, it could substitute YouTubers’ income streams by harvesting their information and experience, probably undermining their complete enterprise mannequin.
Ask YouTube is initially coming to YouTube Premium subscribers, however the firm plans to roll it out to the entire platform sooner or later.
Coding Search solutions on the fly — as a substitute of discovering the best YouTube video
Google launched countless AI options throughout its I/O keynote, however some are extra delicate of their efforts to exchange content material creators’ income streams. Search has some integrations with intriguing capabilities, together with utilizing the brand new Gemini 3.5 Flash’s AI agent capabilities to code easy, fast software program on the fly. For instance, from a fundamental request, it may construct a marriage or travel-planning widget that surfaces related information and deadlines proper within the browser window.
However this so-called agentic coding has one other use case that Google demonstrated within the briefing. Say you requested a query a couple of particular a part of astrophysics, like black holes — to reply, Gemini 3.5 Flash might craft an interactive visible simulation to show how an elaborate idea works. Google calls it “generative UI.”
An instance of “agentic coding” making a simulation of the physics of black holes inside search outcomes.
I might hope that most individuals looking for to know black holes would flip to a good publication with high-quality house and science journalism. However realistically, I would anticipate many people to move to YouTube to see a posh astrophysics idea damaged down visually in a well-produced video. Why would they trouble if Gemini 3.5 Flash whips up a simulation straight from search outcomes?
There’s a whole lot of uncertainty, particularly across the high quality and precision of simulated solutions generated by the brand new Gemini mannequin. But even when they’re sometimes inaccurate, the comfort appears to outweigh their trustworthiness.
Regardless of AI Overviews telling folks to eat rocks again in 2024 and offering different misinformation, folks nonetheless use it as a substitute of scrolling down the web page. The AI startup Oumi, quoted by The New York Times, estimated that AI Overviews utilizing the newer Gemini 3 mannequin have been correct 91% of the time. A handful of well being care organizations and charities informed The Guardian that AI Overviews gave inaccurate strategies for searches on pancreatic most cancers, liver illness and different severe well being circumstances.
I would not be shocked if the explosion of latest AI performance in Search results in fewer folks wanting up solutions in movies. Search cannot substitute streamers (but), however a complete class of YouTube content material — explainers, how-tos and walkthroughs — might see a drop in visitors if viewers do not enterprise past their AI-generated search question outcomes. They usually’ll miss out on all of the curious, bizarre, wild folks making movies to share with the world.
As AI Overviews scale back clicks exterior Google’s ecosystem, YouTube’s conversational AI and Google’s generative search simulations threaten to shrink video content material manufacturing. And that creates a paradox: If creators cease making informative content material as a consequence of an absence of visitors and compensation, AI fashions will not have the datasets they should generate future solutions.
Finally, Google is constructing its empire of “comfort” on the again of uncompensated creators.
