Search engines vs answer engines: Did AI kill Google?

sxsw · Thoughts-

Originally published in Mumbrella.

 

The past few years have seen certain buzzwords get repeated so often that many of us wanted to scratch them out from the dictionary entirely, never to be uttered again. But unlike the ‘unprecedented times’ that came before this ‘new normal’, 2023 sparked a few more powerful phrases that signalled, in many ways, the end of life as we know it. I’m talking, of course, about ChatGPT.

Few words (or letters, technically) have been spoken more at SXSW in Austin than AI. With sessions ranging from film to journalism, not to mention a large percentage of the Brand & Marketing track, everyone has tackled the effects of artificial intelligence in one way or another.

The festival is far from over, but one thing is already very clear: We’re no longer talking about the ‘internet of things’. In fact, it’s no longer the internet of anything. We’re entering the era of ‘assisted computing’ and AI is at the forefront.

The difference between search and answer engines

Whether it’s through your homepage or a widget, most of our interactions with the internet start with search. You type in “how many weeks until Christmas” and a whole universe opens up. So, when search gets unseated as the go-to, the whole ecosystem shifts.

In generative AI (like ChatGPT), you’re seeing the program reviewing the information available (this may be a single text or millions of web pages) and reshuffling this data to give you a response – essentially, it’s an ‘answer engine’.

 

Source: Microsoft

 

But is the answer always right?

When releasing the Trend Report for 2023 at SXSW, Amy Webb said that “computing assistance tools are confidently incorrect”. Based on the information they are fed, the response could be right or wrong – unless you’re running a test where you already know the truth, it’s impossible to know for sure, because it answers your question definitively.

While search engines rely on a similar algorithmic process to display results for your query, the main difference is in the result. Answer engines will, as the name implies, answer your question. But search engines are more likely to give responses to help you to search for the answer you want to settle on. Sure, the much-coveted SERP positions will often highlight a possible answer, but it will be followed by alternative options or even queries.

So, if Google search is already giving us the answers we need, then how can AI possibly kill it?

The demise of search

Well, the answer isn’t as straight-forward as humans being lazy (although let’s never discount that element). There are several factors contributing to the death of search-led internet:

1. Search is fragmented
In 2022, many marketers read the stat by Google saying 40% of Gen Z prefer to search via TikTok and saw this as a warning to join the platform or be dead in the water. But this research wasn’t a call to arms: what it really showed was that we don’t search in just one way anymore.

The internet has become fragmented and we’re more likely to search by vertical – we go to Amazon for consumer products, Dymocks for books or even WebMD for health concerns – than go through one centralised location.

Social media has also given rise to modality-based searches, whether this is reviews on YouTube, local eateries through Maps or even community sourcing on Reddit. How we want to see the results may guide us more than the answer we seek.

2. Ads are everywhere
Are you tired of seeing ads every time you make a simple search? Sure, agreed. However, AdWords still makes up roughly 80% of Google’s revenue, so it’s safe to say they’re not going away any time soon. With consumers increasingly concerned with authenticity and efficiency, it’s not surprising they’re turning to other outlets to get straight to the results.

3. Voice was the next frontier
Do you remember when everyone was saying that voice search would eclipse other types of search by 2022? Well, that prediction wasn’t entirely wrong. While many of us are reliant on our virtual assistants, it seems to be ‘Generation Alpha’ who are ahead in this trend. And why? Because as any parent will attest, kids are far more likely to be able to ask you a question than to type one. And they’ll ask a follow up. And another one…

 

So what about humans?

Generative AI is far from perfect. It is limited by the data at its disposal, it often can’t decipher context or nuance, and even with human controls in place we can still see bias taking shape.

The one thing that we need to continue to cultivate is the human response to AI-generated information. The ability to critically assess the answers provided, to fact check and take them at more than face value, and most importantly, to ask the right questions. These are the skills that will shape our experience of the internet going forward.

So, is Google on its way out? It’s been a hot topic of debate this week and despite all the layoffs, the answer is probably no. They’re working on their own AI answer engine and given the amount of research they’ve shared on the subject they’ll probably be running with the pack in no time.

But as the internet evolves, we need to keep a watchful eye on how it’s changing, and how we’re evolving alongside it. And also that we’re not accidentally starting an AI-pocalypse.