AI Search Events Update: 'New Bing' and Google Bard
AI Search Events Update: 'New Bing' and Google Bard
Is the game changing or is it more of the same?
Both Microsoft and Google held events this week where they debuted new AI features that they’re adding to their respective search engines.
Conversation on social media has been completely hysterical and divorced from reality in the past couple months. Ever since ChatGPT was released, many have been claiming that AI will completely replace search engines and kill organic SEO.
The speculation is officially over. Both Microsoft and Google held in-person AI events this week where they unveiled their plans for integrating generative AI into search results. Is this the death of SEO or just the beginning of a new chapter?
This is a free article. If you’re not already a paid subscriber, upgrade to get access to full-length SEO guides, Q&A sessions, and premium business content not found anywhere else.
SEO isn’t dead
The TLDR is that SEO isn’t dead.
Both Bing and Google will integrate AI chatbots into their search engines, but both will still include organic listings. The AI chatbots will essentially be next-gen featured snippets that appear alongside the normal search results.
I predicted that Google would roll out a generative AI featured snippet earlier last month after Neeva unveiled similar tech for their search engine.
Obviously a small search engine like Neeva has to be aggressive with implementing new features if they want to become relevant. But as you can see from the above screenshot, the tech isn’t fully ready for primetime. It recommended the same product twice (YETI Trailhead), which obviously isn’t great from a user-experience perspective.
But it got them tons of attention, which is exactly what a tiny company like Neeva needs if it wants to survive.
New Bing
Although not as small as Neeva, Bing is an also-ran in the search engine marketplace. Despite Microsoft’s deep pockets, they haven’t been able to make a dent in Big G’s market share (and not from a lack of trying).
Which is why it’s not surprising that they beat Google to the punch with their ChatGPT integration into Bing’s search results.
At their in-person only event, Microsoft unveiled what they’re calling “New Bing” (not sure if their marketing team is unaware of New Coke), which is basically the same as the old Bing but with an OpenAI-powered chatbot next to the search results.
When I tested it out using one of Bing’s publicly-available limited preview queries, it returned a normal search results page with links + an AI chatbot on the right-hand side of the screen.
Hopefully they stick with this layout, because if they do it will be very easy for users to ignore it.
In addition to the normal search results links, there’s a small “Learn more” section underneath the AI-generated content that shows which sites OpenAI generated the answers from.
The so-called “New Bing” hasn’t fully rolled out yet. Right now it’s only available in limited-preview mode, but the company expects to roll it out to “millions” of people in the next few weeks.
I’m currently on the waitlist so I’ll update you guys once I get access.
Google Bard - the $150 Billion error
Google revealed Bard - their conversational AI competitor to ChatGPT - at an event on Wednesday (one day after Microsoft’s event).
Bard is powered by Google’s internal LaMDA AI model, which is distinct from OpenAI’s GPT model.
Similar to “New Bing”, Google Bard will be integrated directly into the SERPs, much like a featured snippet. Organic results will still exist.
Google’s announcement turned into a bit of a debacle (that’s putting it lightly) when Bard hallucinated a fake answer to a query in the official promo video.
The error caused Google’s stock to decline 8% within 24 hours, wiping off $150Bn from their market cap.
It seems like Google let the pressure from Bing get to them, which caused them to jump the gun and release a product that isn’t quite ready for prime time.
A Google spokesman put out the following statement after the disastrous launch:
“This highlights the importance of a rigorous testing process, something that we’re kicking off this week with our Trusted Tester program.” They also said: “We’ll combine external feedback with our own internal testing to make sure Bard’s responses meet a high bar for quality, safety and groundedness in real-world information.”
Detect AI-generated content using Originality.AI, the most accurate AI detection tool available today.
Originality.AI can ID content created by all of the leading generative AI platforms including ChatGPT and Jasper. This tool is a must-have if you have freelance writers or guest posters publishing content on your site.
What do generative AI search features mean for SEO?
The products still haven’t officially rolled out, so it’s still too soon to make definitive predictions.
But based on what I’ve seen so far, I don’t think this will present a major threat to organic SEO.
Contrary to the hysterical predictions about “AI destroying search”, both Google and Bing are still including organic results for all queries.
If the AI chatbots catch on then sites will definitely receive lower amounts of organic traffic. But the number won’t go to zero. SEOs will find a way to adapt and overcome (just like we did with featured snippets).
But there are a few reasons why I don’t think users will start using the tech in a big way.
The first is that annoying chatbot popups on websites have trained users to close them out or ignore them. This creates an effect similar to “banner blindness” where people tune out the chatbots and focus on what they came to the website for.
If people have a similar response to Google Bard and New Bing, then it’s possible that they won’t even notice the chatbots and will scroll down to the organic listings like they’re accustomed to.
The hallucination problem is another serious issue. If Google can’t even get their chatbot to function correctly in the official launch promo video, then it’s doubtful that it’ll work well when actually used by consumers.
Not only does this lower trust, but it could also create legal/regulatory headaches for Microsoft and Google. Imagine if someone asks Bard/ChatGPT a medical question and gets information that worsens their condition or kills them.
It’ll only take a few high profile examples of ChatGPT telling someone that Tylenol is a cure for cancer or people knocking up their girlfriends because Bard told them that the girl can’t get pregnant if she’s on top before regulators/lawyers swoop in and shut it down.
Another legal risk is the fact that both of these tools aren’t truly AI, meaning that they aren’t actually thinking of an answer like a human does. They’re highly advanced scraper tools that are stealing original content from other people. It’s almost certain that there will be a legal challenge against the mass content scraping that these tools are engaged in. Also the EU is notoriously strict about “protecting the little guy”, and may intervene with regulations on generative AI use.
Conclusion
My overall impression of the two AI events is that both companies are scrambling to take advantage of a social media hype train as a marketing play.
I don’t think either company was prepared for generative AI to go viral so soon. They rushed their AI chatbot search products out not because they think they’re ready for primetime, but because they want to stay relevant on social media.
The worst case scenario is that all websites get less organic traffic if people start trusting and relying on these tools without clicking through to organic listings.
There’s a 0% chance that New Bing and Google Bard kill organic search because both Microsoft and Google are including website listings alongside their chatbot answers. There will always be people who want to click through to get a better/more trustworthy answer. Featured snippets in their current form don’t completely kill organic clicks and neither will AI chatbots.
The best case scenario is that the tools don’t catch on. People don’t trust them or they get sued/regulated out of existence. Everyone moves on to the next hype train and forgets all about the “AI will terk err jerbs” hysteria, just like they forgot about all the other tech apocalypse hype trains from the past.
The same people who are losing their shit over “AI destroying everything” are the same ones who hyped up the following tech apocalypse fairy tales:
WFH forever killing in-person work
blockchain replacing everything
voice search killing organic search
NFTs replacing physical art
retail stores getting destroyed by e-commerce
It’s reasonable to assume that the AI hysteria will end up in the same sci-fi apocalypse graveyard as all the other tech hype trains from the past.
Either way, we’ll find out more within the next year.
Stick to the plan and keep at it because SEO isn’t dead yet.
Overall agree with you for one main reason: Incentives.
Google is an Ad company, not a search company. It only cares about giving users answers so they come back again and again to click on those ad links. Even if AI magically is 100% right every time... Google STILL would never only just feature AI results as that would mean there'd be no links... and thus no ad links. Their incentives are aligned with always showing other site links so they can jam ads in there and make money
That being said there's a few points that seem a bit off here
- "these tools aren’t truly AI" - Yes they are. Get what you're trying to say about scraping, but they are 100% AI algorithms
- Annoying chat pop ups are indeed annoying and useless 99% of the time... but chatGPT is materially different, especially for those who know how to use it. It's actually useful (even with the errors). If the chat pop up things begin to match its intelligence levels then that stigma could quickly shift
- AI isn't a fad or going anywhere. If anything, it's iterating at a rate faster than even Moores Law. Training and Inference costs are declining at around 70% p.a. Every 2 years it's 10x cheaper to use (or can be 10x more powerful for same cost). Will it take ALL jobs? Of course not. But it's taking millions every year already and is only getting faster and cheaper according to all the data
Probably the most level-headed take I came across on the topic. Very useful article, thanks.