Google’s not asleep at the wheel. And that’s exactly what makes this moment so blooming interesting. OpenAI has lit the AI fire with ChatGPT, Microsoft is pouring petrol on it with a multibillion-dollar deal, and the one company sitting on the biggest pile of data and infrastructure is suddenly looking…
slow, cautious, and oddly vulnerable.
ChatGPT has been part of almost every business content conversation for a few months now, and yet…
The Question is: Where’s Google?
If you zoom out for a second, it’s nuts to imagine Google being “caught off guard” by a chat interface that predicts the next word in a sentence. The search giant has been publishing ‘large language model’ research for years and have a tonne of clever algorithms in place already. They built LaMDA, and they index more of the open web than anyone else on the planet.
Oh, and they don’t have to go out and scrape content. They already did!
But here we are in early 2023 and the product that’s captured the world’s imagination is not a Google product – it’s ChatGPT, running on Microsoft’s Azure infrastructure, with “Powered by OpenAI” all over the press.
I bet ChatGPT has come up in many of your conversations recently?
OpenAI have spent an absolute fortune building their data. Their ‘model’. Google already has the data, the engineers, the data farms and the search monopoly, but they don’t have the viral moment, and that’s new territory for them.
OpenAI, Microsoft and the Early Land Grab
While Google has been talking cautiously about “responsible AI” (and hat tip to them for that – I’m pleased someone out their is thinking responsibly), Microsoft has gone all-in on the theatre.
Last month (January 2023) Microsoft announced the “third phase” of its OpenAI partnership – a multi-year, multibillion dollar investment, with Azure as OpenAI’s exclusive cloud provider. So basically they are giving OpenAI money to spend on Microsoft products. Odd? Yes. I agree.
The plan is clear: use OpenAI’s models to power Bing, GitHub Copilot, Office, and whatever else can be sprinkled with “AI-powered experiences“.
Suddenly Bing, the butt of search jokes for a decade, is being talked about as the test bed for ChatGPT-style search.
Microsoft doesn’t need to dominate the whole of search to win here. If they can claw back a meaningful chunk of queries (they get a tiny amount compared to Google), especially the high-value commercial ones, that’s billions in play. Google knows that, which is why reports of “code red” inside the company don’t feel exaggerated.
Google’s Dilemma: Move Fast, Break Trust?
On paper, Google should be the one charging ahead. They already have:
- LaMDA, a powerful conversational model (that’s the ‘Chat’ interface) they demoed back in 2021
- A search engine that could instantly distribute any new AI feature to billions of users
- A mountain of behavioural data across Search, Maps, YouTube, Android, Gmail, Chrome, and more
The problem isn’t capability, it’s risk.
If Google bolts a hallucinating chatbot directly onto search and it spits out dangerous medical advice, financial nonsense, or defamatory rubbish at scale, regulators, advertisers and the press come straight for them.
OpenAI can get away with being “experimental”.
Google can’t.
Google’s entire business is built on a perception of reliability: you type, you get something broadly useful, ads show, money flows.
So instead of launching a wild-eyed “ChatGPT for Search“, they’ve announced Bard as an “experimental conversational AI service”, powered by LaMDA, with a limited rollout to “trusted testers” and lots of careful language about responsibility. Even the way they’re framing Bard is cautious: a “collaborative AI service” that can be used alongside Search, not a replacement.
That caution already has a cost. Bard’s first big public outing included an incorrect answer in a promotional demo and wiped around $100 billion off Alphabet’s value in a day. When a single inaccurate fact in a marketing video can erase that much shareholder value, you understand why they’ve been sat on this technology.
So What’s Google Going to Do?
A few things feel almost inevitable.
Google won’t sit back and let OpenAI own “AI answers”
They’ve already revealed Bard and explicitly said it will be integrated into Search. That tells you the direction of travel: expect more query-answer experiences at the top of results, powered by their own models, not OpenAI’s.
They’ll wrap AI in the Google ecosystem, not ship a standalone toy
ChatGPT today is a separate destination. You go to it, like you go to a website.
Google’s power is that it sits between people and everything else.
The logical move isn’t to clone ChatGPT as a separate thing but to sneak AI into:
- Search (summaries, conversational follow-ups)
- Docs, Gmail and Workspace (drafting, summarising, replying)
- Android (assistant that finally feels smart instead of clunky)
They’ll lean hard on fresh data and the live web
ChatGPT has impressed people while being mostly trained on data that stops sometime in 2021. Google’s pitch is that it “draws on information from the web” to provide fresh up-to-date responses, which is a subtle but important difference. If they can reliably blend live search results with model reasoning, they undercut one of ChatGPT’s current weaknesses.
They will try to look “responsible” rather than “reckless”
Expect a lot of talk about safety, guardrails, and encouraging users to “Google it” to verify Bard’s answers. They’re even building that back into the interface: a “Google It” button under Bard’s responses to push people back to traditional results. That’s not an accident, it’s a legal and reputational shield.
What About the Other Players?
We are still in very early days.
Microsoft is clearly hitching itself to OpenAI for the short term. They get differentiation for Bing and a story for investors (“we’re the AI company now”), and OpenAI gets scale, GPUs, and distribution. Whether Microsoft later decides to build or buy its own foundation models is anyone’s guess, but the marriage of convenience is very real.
OpenAI is enjoying its moment as the scrappy outsider, but its dependency on Microsoft’s cloud and capital is obvious. Without Azure, training and serving these models at scale becomes a lot harder.
Apple’s the very quiet one. There’s no splashy ChatGPT competitor (yet), but it’d be naive to assume they’re not building on-device AI features and thinking about how this fits into Siri, privacy marketing, and their hardware ecosystem.
Others (Anthropic, Cohere, Baidu, etc.) are circling, but for now the story is dominated by the triangle of Google, Microsoft, and OpenAI. Baidu is already talking about its own ChatGPT-style bot, which will put pressure on Google in markets where it’s used to being the technical leader, even if not the dominant search brand.
There’s no way in hell Google just accepts a future where “AI assistant” means “Microsoft + OpenAI” in the public mind. They’ll spend whatever it takes (and they can spend without a begging bowl), throw as many engineers at the problem as needed, and use their control of distribution to stay in front.
If You Care About Search and Content, What Matters?
Writing this as of February 2023, from the perspective of a small business and content person, a few practical realities are already clear.
AI is not replacing search overnight. Google is actively designing their own AI system as a companion to Search, not a full replacement, largely because they can’t risk burning down the ad model that pays for everything.
The fuel for all of these systems is still the open web. Google’s advantage comes from years of crawling, indexing and ranking the very content that site owners publish. OpenAI’s advantage comes from scraping a frightening amount of that same content into training sets.
Nobody gets to play this game without data.
The people who keep publishing clear, useful, well-structured content on their own sites will still matter.
Whether the user sees ten blue links, an AI generated paragraph with citations, or a ChatGPT-style answer, there has to be something underneath it to train on and to link back to.
So Where’s Google in All This?
Right where you’d expect them to be:
- Slightly late to the party in terms of PR
- Ludicrously over-equipped in terms of data and infrastructure
- Terrified of breaking the money machine that is Search
- Quietly wiring AI deeper into everything they already control
OpenAI has had the first big moment.
Microsoft has grabbed the hype.
But if you’re betting on who still controls what most people see on the internet in a few years’ time, it would be pretty brave to bet against the company that already owns the front door.
And that company is still Google.