AI's rapid development and integration have been a hot topic for us over the past year, prompting our research on AI companion chatbots and sparking numerous conversations among our global network of content creators. We've made a concerted effort to consider all sides of the AI debate, examining its various forms and applications from AI chatbots to machine learning algorithms. And as we've watched its influence grow, one thing's become clear: AI is everywhere. So with its rapid integration into so many products and services, it’s natural that Google combined it with Search. 

‘AI Overview’ (initially being rolled out in the US), pulls content from existing websites to create a quick summary of information at the top of the Search page. This means users don’t need to spend much time browsing; the details they need are placed in front of them. But, controversially, Google has not allowed users to opt-out, resulting in exposure to an AI service without tick-box consent. 

So, while AI Overview holds water as a concept, it generates some new questions within the AI debate. Let’s explore what these are.

Stealing or Borrowing?

One of the hot topics surrounding the Generative AI debate is the use of data to train models which then create content built on 'learnings' from that data set. The idea that AI developers are using people’s original work to train tools to generate content is a topic that we have already seen spark debate across the industry. AI Overview takes this a step further; the responses it supplies aren’t original, or even purely based on the training data set, but are regurgitated from data pulled from other websites and moulded into something else. 

It causes concern for the value of original content – particularly monetization from views and clicks – potentially putting creators on the back burner while search engine giants utilise the information they’ve produced. And what about websites that pay for content, like VoiceBox? How should we view Google taking ‘the best bits’ and offering it for free? An argument could be made that it’s ok to sacrifice some rights over content for the sake of ranking on search engines, but there’s also the question of how AI Overview will affect page views over time. And if the clicks are no longer coming in, what’s the benefit? We’ve already seen a similar point of contention develop ten years ago with the introduction of ‘Featured Snippets’ which caused an increase in ‘zero-click searches’. It’s a murky, ethical grey area that we expect to become a commonplace question as AI Overview develops. 

Google may only be ‘borrowing’ content, but it creates a power imbalance that risks exploiting the work of web developers, copywriters, journalists, and more. While AI Overview does credit sources, we expect users – in our ever-increasing search for convenience – will be unlikely to click and read the information at its roots. It begs the question: can creators be recognised when we no longer need to leave Google’s search page for answers?

When is Search, Search?

We know that Google is a company that prides itself on providing the best search results for users’ queries, and we can see the benefit of highlighting an AI-generated collective of relevant information at the top of their results page. But AI Overview is moving Google into new territories. Traditionally, Google has focused on curating and ranking existing web content, acting as an intermediary rather than a content creator. This approach meant that Google was not responsible for the content on the web pages it ranked. But, because AI Overview summarises and delivers information, it raises questions about the responsibility it holds for the material it recommends. 

This is a particular worry considering that AI Overview has had some issues fact-checking and understanding satire – meaning it has fed users false, and sometimes dangerous, information, including incorrect advice on what to do if you are bitten by a rattlesnake. This situation echoes a recurring question in the AI field, one that has been highlighted by previous missteps of other AI chatbots and systems. Even with disclaimers warning users not to trust everything the AI chatbot produces, to what extent do the companies behind AI chatbot technology hold responsibility?

Impact on Search Discernment

The integration of AI-generated overviews has now blurred the boundaries between traditional search functionality and the use of generative AI. As ‘Search’ evolves, we anticipate that user discernment about their search results will drop. We appreciate the convenience of AI Overview; it allows users to spend less time searching for the answers and offers bite-size information often hidden on a web page. But we must also consider the argument that by highlighting certain information at the top, Google is potentially encouraging a one-dimensional form of intake, coaxing users into relying on what’s placed in front of them over critical thinking and research. Perhaps it is still the user’s responsibility to check what's presented to them beyond AI Overview. But while plenty have the media literacy to think critically, we suspect that some will become reliant on the answers displayed at the top of search results, especially if those results are seemingly presented by a seasoned tech giant. It raises important questions about the balance between convenience and the erosion of deeper research skills.

A Happy Medium?

We welcome Google’s pioneering approach to AI and understand the drive to develop new technologies. Young people are often early adopters, with many excited about emerging Web 3.0 services. 

It is, however, fair to question whether tech companies are getting the balance right between staying ahead of competitors and having a robust, safe product. We would advocate for greater user information and education given the wider societal impacts across all age groups and sectors. As AI Overview continues to roll out worldwide how can we foster balanced views and critical thinking, especially when AI-generated content may present conflicting or incorrect information? 

Perhaps a happy medium is achievable here. Google should allow users to opt out of AI Overview, restoring their autonomy and delivering the increasingly urgent need for digital rights and fairness. We would also urge Google – and other tech firms – to double, triple, and quadruple-test their AI services before launching them. Too often, we see people exposed to unreliable and sometimes harmful technology that hasn’t had the correct due diligence. Children and young people, in particular, are vulnerable to its effects.  

Finally, there needs to be a bigger focus on AI literacy across big tech and educators. Involving young people in this cross-sector collaboration (who have helped shape digital services such as texting and social media in the past) is one step forward to reducing the ‘top-down’ approach often associated with the legislation that impacts them.