“[Artificial Intelligence, or AI] will change the way people work, learn, travel, get health care, and communicate with each other […] Businesses will distinguish themselves by how well they use it.”
For anyone still in the dark, ChatGPT is essentially a very powerful AI chatbot. Using machine learning to trigger fast responses to text prompts, it can answer questions, write code, produce copy, come up with recipes, analyze data, and generate ideas for anything and everything you could imagine.
Want the inside scoop on the most game-changing technology around? Here’s everything you need to know about ChatGPT, including:
- How are people actually using ChatGPT?
- Which industries are using it most?
- What impact will ChatGPT have on search engines?
- What are people’s concerns around ChatGPT?
- Would people be willing to pay for a premium version of ChatGPT?
- Can ChatGPT hold a conversation?
1. ChatGPT curiosity is high and everyday use cases are varied
At the time of writing, 16% of consumers had used ChatGPT, but curiosity spanned far wider, with 59% interested in using it.
Once people use this AI tool, they tend to return to it. A third of users log on daily – for reference, that compares to around 70% who use social media apps at least once a day.
So why are people loving ChatGPT? 86% of its users find it helpful, and the versatility of the program allows it to be applied to a wide range of use cases.
A huge advantage of ChatGPT is that it benefits from reinforcement learning (a type of machine learning that allows an AI model to learn and improve through trial and error). Over time, this training has made the tool increasingly accurate and reliable to use.
2. IT and software industries are leading the way
AI tools have been voguish for a while. ChatGPT’s been used in the workplace as a way of developing ideas, summarizing blocks of text, and asking for advice.
53% of workers say their company is already using or has an interest in using AI.
This might be simpler technology than what powers ChatGPT – it could be a “classic” chatbot with predefined rules and responses – but it shows how ready companies are to deploy it.
It’s unsurprising that IT and software are the first sectors to use ChatGPT – they’re always on the front line of new technology. But the fact it’s also being used in other industries like insurance and management consulting shows how the technology can have even broader applications for data science.
Here are just a few examples of sectors dipping their toes into the world of AI:
- The travel industry is using ChatGPT for suggestions of activities to send on to clients.
- Real estate companies are using the program to write property listings before sending them out.
- Physicians are also testing the tool’s ability to diagnose patients, and reporting “eerily good” responses. It’s even passing medical exams.
3. Online search is set to change drastically
Search engines haven’t changed much in the past 25 years. You type in a query, and you get ten blue links. Other things have been added over time, but the core user experience has largely stayed the same.
But ChatGPT users are at the forefront of a new way to gather information, with half its users saying they’ve logged on to find information or facts.
68% of ChatGPT users would consistently use it over a search engine to answer questions.
Some important context here: search was already being challenged elsewhere, particularly by younger consumers. Gen Z go to social media to find information about brands, more so than using search engines. Instagram is more popular with US Gen Zs than Google.
Currently, finding information online can be a long process which often involves looking through multiple links for the answers we need. Chatbots powered by a large language model (LLM), on the other hand, can deliver fluent and immediate answers to complex queries.
In a world where speed and convenience are becoming the norm, how we search for information might be about to shift in a new direction.
4. Despite excitement, people worry about ChatGPT’s bias, reliability, and privacy
For all ChatGPT’s promise, people still have reservations. Even well before its release, our research consistently showed consumers tended to be more fearful of AI compared to other new tech.
43% of its users don’t think ChatGPT is a reliable tool they can use consistently.
They might have a point. ChatGPT essentially uses deep learning to infer meaning and relationships between words, based on its extensive training data. Fed a massive amount of text from the internet, it can make statistical predictions about words and phrases that should appear next to each other. This is different to other kinds of AI that are “rule-based”, where categories are pre-defined in their programming.
Think of it this way: ChatGPT doesn’t know what a dog is, but based on where dogs appear in its data set, it can infer that it’s a four-legged mammal commonly kept as a pet (at least, that’s how ChatGPT described it to us).
But this means it can’t verify its own output – it can’t tell you if something is a dog or not. That’s one of the reasons why math, for example, isn’t ChatGPT’s strong suit.
The program can also be used for unethical purposes, either by spreading misinformation, or cheating on tests – something 64% of consumers say they’re concerned about when it comes to AI.
There have been cases of students plagiarizing work, and instances where hiring managers have caught candidates using ChatGPT during the job application process. Employers are already finding ways to catch applicants who apply with AI.
Some companies have even taken steps to ban ChatGPT in their workplace due to concerns about third-party software. Prompts that a user puts into ChatGPT may also become a privacy risk, as any sensitive details given could become public domain. And bias is another issue to consider.
46% of users worry that ChatGPT is a biased tool.
As language models like ChatGPT are trained on data pulled from the internet, they can reflect the biases in that original data. It may associate certain concepts more with men than women, for example.
No wonder its validity is in question. Campaigns like “Correct the Internet” are trying to fix these issues, but as people start to use artificial intelligence instead as a way to search, there’ll be new biases to fix.
5. There’s appetite for a premium version
Compared to “regular” search, ChatGPT queries are expensive. And at some point, OpenAI and other providers of AI technology will have to start turning a profit.
Given it’s already being plugged into search engines, it’s likely ad revenue will come into play eventually, though we don’t yet know how consumers will feel about ads appearing in their responses.
The other revenue stream OpenAI is exploring is a premium tier. ChatGPT currently offers a subscription plan, ChatGPT Plus, with benefits like access to the program during peak times, faster responses, and priority access to new features.
But as this is a completely new type of technology, and one that’s been available for free to so many people, would anyone be willing to pay for it? As it turns out, they would.
Almost half of ChatGPT users say they’d be happy to pay for a premium version.
In the future, a premium version of the chatbot could offer capabilities and features which aren’t available to free users, creating even more demand.
But even in these early days, the message is clear. People are interested in a text tool like this, and around half say they’re willing to pay for something with even smarter or more complex features.
6. It’s not human, but it sure feels like it
Interacting with ChatGPT means having a conversation. And even when interacting with an AI, we can’t resist having a good old chat like we would with anyone else thanks to its near-human feedback.
Two-thirds of ChatGPT users found themselves chatting to the bot the way they would to a live person, and half say they couldn’t tell the difference between a ChatGPT response and a real human response.
For marketing and advertising in particular, ChatGPT’s human-like, natural language responses mean that marketers could use the program to create personalized, engaging ad choices and campaign content creation. While output may need tweaking, it might provide more efficient ways of working.
It’s also interesting to think about its use cases in the long term. We’re already seeing applications emerge that use language models for companionship, and more may arrive as the technology becomes more widespread.
The bottom line
This is just a snapshot of how consumers felt about ChatGPT shortly after its release. With the expansion of its API and various plugins, the technology will spread and will likely be layered into most workflows and online behaviors.
But our data reveals vital clues into how real people will actually feel as it becomes more present and popular.
New users will see the benefits and keep returning to it, and they’ll probably turn to artificial intelligence instead of traditional search engines for some of their queries. Many will think the pricing for a premium tier is good value and add it to their list of subscriptions. And some may even use the tool so much they start to build a personal connection with it.
But even as they do, they’ll feel cautious about its safety and how reliable its answers are.
As AI technology continues to evolve, there’ll be more innovative applications and use cases for ChatGPT. Watch this space.