- ‘What is artificial intelligence?’ is the UK’s most Googled AI question at 125,000 searches a month, followed by ‘What is ChatGPT?’ – showing that Brits are still getting to grips with what AI actually is.
- Brits also show interest in AI’s environmental impact, with ‘Is AI bad for the environment?’ and ‘How much energy does AI use?’ making the top ten searches.
- An AI expert answers Brits’ most-Googled questions and explains what they all mean.
You probably use AI every day, whether you realise it or not. But do you know what any of those buzzwords you keep hearing are actually referring to? And from the UK’s search trends, it’s clear lots of people are still working it out.
A new study from Snowfox, an AP automation and invoice-processing AI company, reveals the most common AI-related searches in the UK – and most of them show people are still trying to get comfortable with the basics.
To clear things up, Snowfox.ai CEO Ilkka Lassila takes the UK’s most-searched AI questions and explains them in a way that makes sense.
1. ‘What is artificial intelligence?’
At 125,000 average monthly searches, this question is miles ahead of the others – and to Lassila, that makes sense. “AI kind of exploded into everyday life,” he says. “Most people didn’t get a gradual learning curve. It went from ‘something the tech world mentions’ to ‘something in all your apps’ pretty much overnight.”
So, what is artificial intelligence? Lassila puts it like this: “It’s basically software that can think through tasks the way a human might, recognising patterns, solving problems or making suggestions. Whether it’s recommending a film or helping process invoices, it’s doing work that used to require human judgement.”
And with AI taking on all kinds of tasks in the background, Lassila says it’s only natural that people want to understand the basics. “If something is making decisions for you – recommending things, sorting things, helping you at work – you’re going to want to know what’s actually behind it,” he observes.
2. ‘What is ChatGPT?’
Brits search this question an average of 106,000 times each month, and Lassila says this is down to how quickly ChatGPT became part of everyday life. “For a lot of people, ChatGPT was the first AI tool they actually used,” he says. “You could ask it a question, and it talked back in full sentences. That was new.”
He explains that ChatGPT is a language-based AI that’s learned from a huge amount of text. “It recognises patterns in language, so it can answer questions, write text, explain things, help brainstorm and more. It’s basically predicting the next word over and over again – but in a very sophisticated way.”
According to Lassila, the sheer amount of searches shows people aren’t just using it casually. “Once someone sees what it can do, they naturally want to understand what’s behind it. These searches don’t come from hype – they come from people interacting with the tool and thinking, ‘OK… but how does this actually work?’”
3. ‘What is generative AI?’
This question racks up nearly 60,000 average monthly searches, and with good reason – generative AI is the part of AI people can see and interact with most clearly.
“Generative AI is technology that creates new content rather than just analysing existing data,” Lassila explains. “It can generate text, images, music, designs, voice clips and even video. That’s why it feels so different from older AI systems.”
Tools like ChatGPT, Midjourney and DALL·E are all generative AI models. “They take what they’ve learned from huge datasets and use that knowledge to produce something new,” he adds. “It’s the first time people are seeing AI behave creatively – and that’s why the interest is so high.”
4. ‘How does AI work?’
This question is searched an average of 33,000 times a month – showing it’s something a lot of people are trying to get their heads around.
“Most modern AI works by learning from examples,” Lassila says. “You feed the system large amounts of data – images, text, audio, anything – and it gradually learns patterns. Once it recognises those patterns well enough, it can make predictions or decisions on its own.”
It’s the same principle that lets an email filter detect spam or lets a car recognise road signs. “It’s not magic,” he emphasises. “It’s statistics, pattern recognition and a lot of training cycles. But when you do it on a huge scale, the results can look surprisingly intelligent.”
5. ‘What is deep learning?’
This more technical question still receives over 23,500 average searches a month, and Lassila says that’s because deep learning sits at the heart of most modern AI – even if people don’t realise they’re using it daily.
“Deep learning is a type of machine learning that uses many layers of interconnected nodes – we call them ‘neural networks’ – to recognise patterns,” Lassila explains. “Each layer learns something slightly different: one might pick out shapes in an image, the next recognises those shapes as a face, and another layer identifies the expression on that face.”
Put simply, deep learning works by feeding an AI massive amounts of examples until it becomes very good at spotting subtle details. “It’s why your phone can sort photos by person, why translation apps sound more natural now, and why tools like ChatGPT understand context instead of just keywords,” he says.
6. ‘Is AI bad for the environment?’
Searched an average of 16,450 times a month, this question shows that more people are thinking about the environmental side of AI – and Lassila says that’s a good thing.
“Large AI models use a lot of electricity – especially during training, which can take weeks on powerful hardware,” he says. “It’s understandable that people want to know if that comes with an environmental cost.”
He adds that everyday use, like sending a ChatGPT prompt or generating a short piece of text, uses only a fraction of that energy, but the bigger systems behind the scenes still have an impact. “The industry is moving towards more efficient chips and greener data centres,” he says. “And the fact that people are asking these questions really matters — it pushes companies to keep sustainability on the agenda rather than treating it as an afterthought.”
7. ‘What is a chatbot?’
This still attracts nearly 12,000 average monthly searches, even though most people have encountered chatbots many times – often without realising it.
“A chatbot is simply software designed to hold a conversation,” Lassila explains. “It can answer questions, help with customer service, guide users through processes or provide information automatically.”
Older chatbots followed rigid scripts, but modern ones – often powered by AI – can respond more naturally. “That’s why some people Google this,” he says. “They’re trying to figure out whether they were talking to a person or a bot.”
8. ‘What is reinforcement learning?’
This term appears just over 9,000 times a month, on average, usually from people wanting to understand how advanced AI “learns by doing.”
“Reinforcement learning trains AI through trial and error,” Lassila explains. “The system tries something, gets feedback – a reward or penalty – and uses that information to improve next time.”
It’s the technique that helped AI beat human champions at games like Go and helps robots learn how to walk or grasp objects. “It’s not something most people use directly, but it’s one of the most fascinating parts of AI once you grasp the idea,” he says.
9. ‘How much energy does AI use?’
Brits search this question an average of 7,800 times a month, and Lassila says it reflects a growing realisation that AI isn’t just “digital” – it runs on physical servers that use real electricity.
“AI relies on huge data centres with physical hardware,” he explains. “Training a large model is extremely demanding. That process uses far more electricity than most people expect,” he says. “But once a model is trained, the energy needed to run individual prompts or tasks is surprisingly small.”
10. ‘What is machine learning in AI?’
This question receives 6,100 average monthly searches, and Lassila says it focuses on the foundation that almost all modern AI systems are built on.
“Machine learning is a method where instead of giving a computer strict rules, you teach it by showing examples,” he explains. “If you want an AI to recognise cats, you don’t tell it what a cat is – you show it thousands of pictures and it figures out the patterns itself.”
It’s the idea behind spam filters, fraud detection systems, recommendation engines and more. “Machine learning is basically pattern-spotting at scale,” Lassila says. “The more examples it sees, the better it becomes.”
Top 10 AI questions searched monthly by Britons
