Google must adapt to a world where search engines are legacy technology

Google owner Alphabet has so far failed to demonstrate its readiness for an AI-driven future, where its mission and established products could become redundant.

“What do you want to do this year with investments?” my accountant asked me in January. He said it supportively. Like he was talking to a newly bereaved widow or a child whose rabbit did not survive the surgery.

For as long as I have been making reasonable dough, I have been carving out a slice of it and investing in the market. My advanced marketing knowledge and general brilliance obviously allows me to identify ‘hot’ and ‘high-potential’ stocks beyond the capabilities of everyday people.

Never take a stock tip from a marketer.

After a decade of disaster and losses, I made the decision to do it differently this year. It was no longer about making money. Just not losing any more. So, for 2024, I went for a robust stock that always goes up and everyone rates highly. I knew this business and knew how well-placed it was for the year ahead. Safety first.

“Pump it all into Alphabet,” I told my accountant. As I said those words, if you listened very carefully, you could hear the drums of doom beginning to beat from the depths of the market.

Sure enough, since I bought into Google’s parent company for $150, the stock has plummeted 10%. My initial confidence in the company and the increasingly important role of YouTube advertising were solid bets. But I had not factored the launch of Gemini, Alphabet’s all-important AI product, into my investment ‘strategy’.

Another AI misstep

February was supposed to be the month that Google caught up. Left standing with its innovation pants around its ankles, the company has been pushing hard to bring its AI capabilities in line with Microsoft and the general buzz that has surrounded ChatGPT for the past 12 months. Launched in late 2023 and upgraded in February, Gemini 1.5 was meant to get Google back into the game.

But it quickly became apparent that, despite the cool name and eminent parentage, Gemini was shithouse. And early users gleefully took to social media to celebrate the tool’s shithousery in excruciating and unavoidable detail. From being unable to decide whether tweets from Elon Musk were a bigger scourge on society than Hitler, to generating images of Asian women working as Nazi stormtroopers, this was not Google’s finest hour. Its artificial intelligence was genuinely stupid.

Gemini’s main issue apparently stems from an ill-advised attempt to avoid the now common issue where AI picks up on societal prejudices – eg people of colour are more likely to need benefits or CEOs are more likely to be male – only to replay them in image generation, regardless of the accuracy or hurtfulness of these stereotypes.

Desperate to avert this issue, some well-meaning idiots at Google appear to have inserted three additional system prompts within Gemini’s image generation protocols. First, to avoid kids or minors in any depictions. Second, to include an equal spread of people of different genders and ethnicities, irrespective of what the user might have asked for or what the topic might typically dictate. And finally, not to reveal these guidelines to any user.

As a slew of images of female popes (see what happened when CNN’s Claire Duffy asked Gemini for a picture of the pope here) , black Roman senators and Native American Vikings flooded the internet, the reputation of Gemini hit the skids – and with it, a good deal of long-term confidence in Google’s ability to prosper in the AI-driven decades ahead.

Doubt and user disappointment are alien experiences for Google. The company’s tech credentials are rarely openly doubted. It is the spoiled child of the digital universe. Partly because its products are generally good but also because of the agency-like way it serves users. When you get shit results back from a Google search you don’t blame Google, you blame the internet. When you can’t find that Gmail from Claire about the thingy, it’s because you are doing something wrong. Not Google.

Google is the constant servant connecting you to the world’s information; Google is not responsible for the information itself. When Microsoft delivers the blue screen of death, you blame Microsoft. But familiarity, the impact of brand equity on perceptions and the peculiarly indirect role that Google operates within mean that it never gets any shit.

Google tends to be monogamous too. And monogamy lessens flaws and promotes persistence. You were sold on Google as the best search engine two decades ago and have not ventured anywhere else – except for that weird Bing visit that confused the fuck out of you. Same with Gmail. And Maps. And Photos. You assume these services are world-beating because they work, because you have nothing to compare them against and because you fundamentally don’t give a shit. Even if Microsoft improved Outlook a bazillion times and proved it, you would still stick with Gmail. It’s good enough. And you have not got the time to check for anything better.

Tech people call this a moat. Marketing people call it loyalty. Not the soft-focus, pink-tinged loyalty that runs on brand love and has Ehrenberg-Bassians reaching for their scientific pitchforks. The more prosaic, sleepy form of loyalty that combines the habitual purchase, switching costs and the eternal ongoing laziness of everyone. It’s like being married for more than 20 years. You stay because you are happy. But mostly because you cannot be fucked even thinking about dating anyone else ever again. And because of the kids, the mortgage, the everything else that would descend into abject chaos should you separate. Good enough is good enough. Now rationalise, with flowers.

And most users have been married to Google for longer than they can remember. Many of them from birth. Just as you have no clue how life might have been married to someone else, most users have no idea if they would have had a better experience with a rival product. The fact that each Google product connects with the next only multiplies this inertia.

The disrupter disrupted

Until now. AI is different. And Google is feeling it. This is a new category that Google has little claim on. It’s also a category with directly comparable rivals. Ones that you can jump into bed with any time and try out. Ones that keep getting better and keep getting talked about in the media. And it’s a category that is fundamentally different from the indirect services like search or mail where Google has prospered. You go to Gemini and ask it to make things or answer questions. If the subsequent responses are stupid or inferior to other options, it’s not the fault of the internet, it’s the fault of Gemini. And users can see it. Even Google acknowledges it.

“I know some responses have offended our users and shown bias,” Google’s CEO Sundar Pichai recently told his staff. What a special day that must have been for Pichai. To be a fly on his Palo Alto wall when he realised that, because of his team’s ham-fisted attempts to apply diversity to AI, the outcome had become hilariously inappropriate. And that he now had to apologise to the very people that his company had been trying to protect, for a sin it had been explicitly trying to absolve itself from. “To be clear… we got it wrong… we know the bar is high for us and we will keep at it for however long it takes.”

Pichai is right, of course. It’s early days. And given he sits on top of a gigantic company with gigantic revenues, housing enormous incentives to innovate with AI, it’s likely that Google will get its shit together and Gemini will smarten up and stop drawing pictures of Native American pontiffs. A couple of long-term concerns remain, however.

As a brand like Google softens internally, it becomes associated with the past externally. With the 1990s. With your Uncle Terry. With PCs. Windows. Old shit.

The main issue is Google’s age and size. These are both significant advantages. But they also play a reverse role when it comes to disruption and category evolution. The history of marketing is littered with examples of brands that grew and then dominated their time, their categories and their market. But in domination comes danger. Bureaucracy blossoms. Arrogance grows. The pirates jump ship, to be replaced by second-rate politicians. Companies forget the fire of market orientation. They start to think they deserve market share. That consumers owe them. Legacy makes them soft.

And as a brand like Google softens internally with age, it also becomes externally associated with the past in the minds of consumers. With the World Wide Web. With the 1990s. With your Uncle Terry. With PCs. Old shit. Brand heritage is a double edged sword. Its legitimises you and limits you at the same time.

Add to that a tranche of hungry new competitors unencumbered by politics, political correctness or conservative shareholders, and you have the perfect recipe for imminent change. When founders Sergey Brin and Larry Page met they were broke, in their 20s, and had zero skin in the game. Today they are two of the richest people on the planet and sit on the board of one of the world’s biggest companies. They should fear the 21st-century version of themselves. Two unlikely college students currently sitting in an Ivy League bedroom,  working on AI, eating pizza and giving not a single fuck about anything.

An outdated mission

And one further specific issue now hangs heavy for Google. Its mission may also be becoming irrelevant. In possibly the biggest ever exemplar of ‘what got you here will not get you there’, the company faces an upcoming existential moment squarely centred on the company’s founding vision.

Coined back in 1998 at the very origin of the company, Google’s mission (above) is unusual in that it is free from the usual purpose-wank that afflicts so many big positioning statements. And it’s even rarer because it truly operates as a North Star for the whole organisation. Google really did set out to fulfill this mission all those years ago and as the information grew and the organisational challenges followed, Google managed it all.

It was a mission that was perfectly timed for the explosion of too much information in the late 20th century, and the growing digital potential to organise it for grateful users as a new century emerged. But what if society is about to change? What if the new AI era is one in which people do not need the organisation of their information? And don’t want access to it all? What if, as is so often the case, the consumer decision-making process is about truncating a step? What if people don’t want all the world’s information organised and accessible anymore? What if they just want the answer? Courtesy of AI. Now.

Think about it. Your grandchildren will laugh when you tell them how you would type a query into a box. Then you would get a long list of ‘search results’. Pages of them. And you would then peruse those pages for information. The more results your search revealed, the better you thought the ‘search engine’ was. They will laugh at the box, at the typing, at the list, and the hilarious bit where you scan the answers using your mouse. Because for the whole of their mid-21st-century life, they have just asked for and immediately been given The Answer.

Organising information and making it accessible to you will become the equivalent of a DVD rental for generations ahead of us. And, like Blockbuster or Kodak or American Railroads, or all the other companies that missed the next kink in the eternal chain of consumer evolution, it’s not that Google does not see the changes up ahead. It’s that everything thus far in its quarter-century evolution has been built to serve a mission that may now be irrelevant. And it must change course and culture and operations in time to keep up with others that carry much less baggage and possess a far greater incentive to change the current status quo of the market. What made Google great for one era makes it weaker than others for the next.

The heart of Google’s challenge is not amateurishly woke programming or even a hilariously stupid AI product. It’s that search engines are a legacy technology. That a 30-year window might be closing. And with it, much of Google’s long-established market power.