Listen to this article
This audio is generated using AI
Investors should consider the investment objectives, risks, charges and expenses carefully before investing. This information and other information about the Funds can be found in the prospectus and summary prospectus. For a prospectus and summary prospectus, please visit our website at bailliegifford.com/usmutualfunds Please carefully read the Fund's prospectus and related documents before investing. Securities are offered through Baillie Gifford Funds Services LLC, an affiliate of Baillie Gifford Overseas Ltd and a member of FINRA.
Maya runs a small US travel site. Last year, her Plan a Weekend in… guides paid the bills. This year, she wakes to a strange pattern: website crawls are up, but click-throughs are down. AI answer engines skim her pages to construct a reply inside their users’ chat boxes, so fewer people ever reach her site. She still pays to create content and host the website. The audience just doesn’t arrive anymore.
Speaking to Matthew Prince, chief executive of Cloudflare, this summer provides context: Compared with the ‘Old Google’ search model, a site visit is about 750 times less likely using OpenAI and about 30,000 times less likely via Anthropic. The bargain of “we let search engines crawl, you send us readers” has cracked.
What’s breaking
For 25 years, the web ran on a simple swap: ‘crawl → click’. AI answer engines are rewiring it to ‘crawl → answer’, often with no visit at all. More and more of the internet is machine traffic. This is muddying measurement, ad spend and trust. Analytics are noisy, and revenue patterns are changing.
About 20 per cent of the web’s traffic goes through Cloudflare’s network. It is responding to this shift by blocking many AI crawlers by default and piloting a way for sites to charge per crawl so creators aren’t left footing the bill. That’s an early sketch of a new deal.
So what might the AI-era internet look like?
1. Pay what you use
If an AI system like OpenAI’s ChatGPT uses Maya’s work to craft an answer, it could pay for that use, just as a cafe pays for the electricity that keeps the lights on.
With Cloudflare’s pilot, when a bot asks for a page, Maya’s server can reply with “payment required” and a tiny price. If the bot pays, it gets the content. If not, it doesn’t. That keeps instant answers for users but restores a fairer exchange for creators.
Casimiro - stock.adobe.com
Stablecoins – cryptocurrencies designed to have a stable value by being pegged to a fiat currency, commodity or other financial asset – could be how the money moves, outperforming traditional credit card networks. Cards weren’t built for fractions of a penny moving machine‑to‑machine. Stablecoins make small, instant payments possible.
Circle, the company behind the USDC stablecoin, and others have outlined a blockchain-based internet protocol, x402. The intent is for AI agents to pay these micro‑fees in USDC and other stablecoins automatically and get back to work in seconds. Cloudflare is partnering with Coinbase to encourage the adoption of this protocol. This model keeps the convenience of instant answers and adds a ‘meter’ so value flows back to the source.
This is a democratic vision. An alternative is that the AI answer platform pays creators a monthly sum based on their content’s engagement. Here, value accrues to the demand aggregators – the AI crawler – in a similar way to Google and search. Alternatively, platforms strike licensing deals (eg with Reddit and The New York Times).
2. Provenance: show where it came from
AI-generated content is exploding. Therefore, origin increasingly matters. Maya can attach Content Credentials to her work. These tamper‑evident labels say who made a photo or article and how it was edited.
The Coalition for Content Provenance and Authenticity (C2PA)’s open technical standard is being baked into infrastructure so those labels survive across the web. Providers now preserve Content Credentials through their image pipelines, helping original work stand out. Combining this with blockchain technology to record this metadata creates an immutable record.
Provenance also needs people, not just pixels. Privacy-preserving tools like Private Access Tokens (used in Apple’s ecosystem) and Cloudflare’s Turnstile help services tell real visitors from bot swarms without intrusive tracking.
When Maya turns these on, her audience quality improves, and advertisers on her site have more confidence in what they’re paying for.
3. Accountability: hold AI agents to a standard
As AI agents learn to browse, summarise and act, large enterprises are likely to insist on guardrails. That means saying what an agent is allowed to fetch and why – recording what it actually did – and enforcing consequences when it ignores rules, including payment rules. For Maya, who hosts her site on Amazon Web Services’ cloud, this turns ‘black boxes’ into ‘glass boxes’ that are inspectable and correctable. AI answer engines that play by these rules will be easier to trust and do business with.
The future: agentic commerce?
This all leads to agentic commerce, which could accelerate Maya’s business. Instead of waiting for people to click her website, Maya lets AI helpers find and buy her trips directly. A shopper tells their AI: “Plan a 10‑day Croatia trip for two under $4,000.” The AI then pulls up Maya’s pre-made package, asks her bot to tweak dates or hotels, and then pays, all inside the AI chat box.
Maya runs her business on Shopify’s ecommerce platform, and its software makes Maya’s trips easy for AIs to read and buy, allowing her to set simple rules. For example, what can be swapped, how much can be discounted, etc. For payment, the AI can again use stablecoins like USDC to split money instantly between Maya and Croatian suppliers, cutting card fees and currency hassles.
Potential losers in this vision are free-search-dependent sites – is this the end of the ‘listicle’? – non-paying AI aggregators and bot-ridden social platforms. Advertising may also keep shifting towards trusted walled gardens and away from the open internet.
Meanwhile, datacentre and power suppliers benefit regardless of who wins, and emerging compute marketplaces could ease capacity and cost pressures.
Back to Maya
In the future, Maya opens her dashboard. Crawls are up, but now a meter ticks. Answers pay a sliver, credentials travel with her assets, and agentic commerce converts guides to bookings.
Perhaps this is how AI breaks and rebuilds the internet.
Curious about generative AI? Watch how Runway put investment manager Kirsty Gibson in the movies.
AI generated images created with Runway.
Risk factors
The Funds are distributed by Baillie Gifford Funds Services LLC. Baillie Gifford Funds Services LLC is registered as a broker-dealer with the SEC, a member of FINRA and is an affiliate of Baillie Gifford Overseas Limited. All information is sourced from Baillie Gifford & Co unless otherwise stated.
As with all mutual funds, the value of an investment in the Fund could decline, so you could lose money. International investing involves special risks, which include changes in currency rates, foreign taxation and differences in auditing standards and securities regulations, political uncertainty and greater volatility. These risks are even greater when investing in emerging markets. Security prices in emerging markets can be significantly more volatile than in the more developed nations of the world, reflecting the greater uncertainties of investing in less established markets and economies. Currency risk includes the risk that the foreign currencies in which a Fund’s investments are traded, in which a Fund receives income, or in which a Fund has taken a position, will decline in value relative to the U.S. dollar. Hedging against a decline in the value of currency does not eliminate fluctuations in the prices of portfolio securities or prevent losses if the prices of such securities decline. In addition, hedging a foreign currency can have a negative effect on performance if the U.S. dollar declines in value relative to that currency, or if the currency hedging is otherwise ineffective.
For more information about these and other risks of an investment in the Funds, see "Principal Investment Risks" and "Additional Investment Strategies" in the prospectus. There can be no assurance that the Funds will achieve their investment objectives.
This communication was produced and approved in October 2025 and has not been updated subsequently. It represents views held at the time of writing and may not reflect current thinking.
This communication contains information on investments which does not constitute independent research. Accordingly, it is not subject to the protections afforded to independent research and Baillie Gifford and its staff may have dealt in the investments concerned.
As at October 2025, Baillie Gifford held Cloudflare and Shopify. A full list of holdings is available on request and is subject to change.
176413 10058289
