AI Overviews, announced by Google on the 14 May, has been rolled out across the US and is beginning to roll out globally. As the rollout is ongoing, there is a lot of discussion about whether it is ready for full launch. Many SEOs are even opting to find ways to prevent their clients’ website information from appearing within AI Overview snippets.
If you haven’t seen AI Overviews yet, it provides snippets which feature answers crafted by Google’s AI Gemini that take information from a variety of internet sources to provide a “straight answer” to your query. This is provided at the top of the search engine results page.
Social Media Response
On almost all social media platforms, there have been comical but widely spread instances of misinformation noted by both search optimisation specialists and the general public.
An example from one X user Peter Yang included telling users they should add glue to pizza sauce to give it more tackiness.
Another example tells searchers to use “blinker fluid”, a common internet joke which does not actually exist.
Such misinformation likely stems from Google training Gemini with data from platforms like Reddit, which includes both factual and satirical responses. This highlights AI's current inability to differentiate between genuine and satirical information.
There have been strong responses on almost all platforms, whether its TikTok, LinkedIn or X.
In response to some of these wild responses, Google’s spokeswoman Lara Levin said in a statement: "Many of the examples we’ve seen have been uncommon queries, and we’ve also seen examples that were doctored or that we couldn’t reproduce."
She also replied: “The company (Google) will use ‘isolated examples’ of problematic answers to refine its system.”
Could There be a Legal Response?
Since AI has gained popularity and developed rapidly over the last few years, governments have taken measures to put in place legislation on responsible AI usage.
There is a question over the need for a debate on whether Google is being “responsible” in their use of AI Overviews.
In addition to the potential backlash over their perceived lack of responsibility, they could also see pushback from governments over the spread of misinformation due to some of the responses AI Overviews provides.
The 1996 Communications Decency Act, section 230 in the USA, widely protects Google from being liable for third-party content they display, but there is a question over whether this includes AI generated responses provided by the platform.
Google controls around 83.49% of the search engine market of according to Statista. Even getting a small percentage of these answers wrong when there are 8.5 billion searches a day can result in a significant amount of misinformation.
How are SEOs Choosing to Proceed?
SEOs very quickly realised that this could be an issue, especially within certain industries, such as the Your Money Your Life (YMYL) sector. SEOs have discovered that using the nosnippet feature works to prevent their clients’ websites from being shown within AI Overviews responses.
In response to Glenn Gabe, Google’s John Mueller confirmed on LinkedIn that this should work, it may just take longer than other features provided by Google that are available to SEOs and webmasters as it is still new and being implemented.
At present, most reputable SEOs are reminding people to not be too hasty in changing their strategies with this update, and to wait and see the effect this will have on website and content performance.
Personal Opinion
I personally think in some industries, it is imperative at the moment to use nosnippets to avoid AI Overviews until Google has refined the tool. This is especially true in heavily regulated industries, such as gambling and finance, where misinformation could have detrimental effects for the brand. It is also important to keep an eye on whether your brand is being showcased in AI Overview snippets.
AI Overviews is just not ready for large-scale rollout yet. This could lead to the market share for Google decreasing in favour of other search engines, such as Bing or Yahoo, until the misinformation is resolved or users have the choice to “opt out”, which they currently do not.
