IT Pro is supported by its audience. When you purchase through links on our site, we may earn an affiliate commission. Learn more

What is ChatGPT and what does it mean for businesses?

ChatGPT has taken the industry by storm, opening up a realm of new possibilities in fields from healthcare to publishing

ChatGPT on a smartphone in front of a display that shows AI embedded on a chip

Generative artificial intelligence (AI) has been grabbing headlines recently for its supposed ability to cut costs and time constraints across various industries. One of the most widely used tools in generative AI is OpenAI’s ChatGPT – an extension of its GPT-3 language model that was made available to the public in November last year.

With the market set to expand to $109 billion US (approximately £91.4 billion) by 2030 according to Grand View Research, these tools are becoming more and more important in the enterprise landscape. Since ChatGPT launched, a number of businesses have announced they’ll explore using it in their tools and services, including the online publication BuzzFeed. It’s also been used to pass various exams at a degree level, raising questions for the future of learning and training. 

What is ChatGPT?

ChatGPT is a chatbot powered by OpenAI’s GPT-3.5 language model. The company calls ChatGPT a “sibling” of their InstructGPT model, which is the default language for OpenAI’s API.  

ChatGPT’s model was trained using a combination of human feedback and InstructGPT’s data. From there, the model was rewarded by a process that integrated what Open AI calls human AI trainers. That process, from the organisation’s perspective, was about “fine-tuning” and removing bias and nonsensical responses.

“We needed to collect comparison data, which consisted of two or more model responses ranked by quality,” OpenAI says. “To collect this data, we took conversations that AI trainers had with the chatbot. We randomly selected a model-written message, sampled several alternative completions, and had AI trainers rank them. Using these reward models, we can fine-tune the model using Proximal Policy Optimization. “ 

At this stage, at least from a privacy and ethics perspective, ChatGPT is asking more questions than it’s answering – even if its job is to answer questions. It should be noted this model is still only in the research release stage and that Open AI has a long history of powering these types of tools. Github’s co-pilot tool, for example, is powered by OpenAI. 

What are the potential use cases for ChatGPT?

If you ask tech entrepreneurs then you might get the answer you’d expect, that ChatGPT’s language model can reinvent every industry under the sun. If you ask pessimists, many of who work in education, it’s a reason to run to proverbial panic stations. Many, like Vice’s Edward Ongweso Jr or internet essayist Evan Puschak, strike a middle ground. 

After gaining access, you can send it a prompt and it will return an AI-powered passage of text. IT Pro has used it to produce some astoundingly bad Christmas cards, for example, but its uses are more widespread, and serious. Content creation, as in the case of BuzzFeed, is one avenue, as is customer service or even corporate research.

For one interesting example of an area in which ChatGPT can flourish, we can look at research coming out of Drexel University. A study released in December 2022 found the same natural language processing (NLP) techniques ChatGPT uses can be deployed to identify Alzheimer’s patients. 

When the same model ChatGPT is using was trained on data from Alzheimer’s patients’ interviews, it more accurately identified and predicted Alzheimer’s patients’ test scores. The next step for the researchers, according to a press release from when the study was published, is to work towards developing a digital tool to allow medical settings like doctor’s offices to screen for Alzheimer’s. This demonstrates the sheer variety of potential use cases for a tool like ChatGPT.

How much does ChatGPT cost to run?

Although OpenAI hasn’t publicly revealed its costs per query, industry experts have estimated the tool costs millions of dollars per month to run. 

Related Resource

Solve global challenges with machine learning

Tackling our world's hardest problems with ML

Whitepaper cover with title over image of couple holding hands in the Artic, looking out towards the ice and waterFree Download

While most of the use cases you’ll hear can safely be characterised as silly, like imitating Jane Austen’s writing style, there is no shortage of money being thrown Open AI’s way. 

At CES 2023, Microsoft announced a $10 billion investment into OpenAI, the purpose of which will be to integrate ChatGPT into Office and the Bing search engine. According to Semafor, that would bring OpenAI’s market valuation to an eye-watering $29 billion. 

When that deal is done – the structure involves Microsoft recouping its investment and taking 75% of OpenAI’s revenue until they’ve done so – Microsoft will own 49% of the company. For context, Reuters found OpenAI is expecting in or around $200 million in revenue in 2023, with that number rising to $1 billion in 2024.

Why is ChatGPT so controversial?

Alarm bells are ringing across various industries – from the creative sector to academia – given the capacity for tools like ChatGPT and MidJourney to mimic human creativity. Academics and researchers, in particular, are worrying about the prospect of plagiarism – which is a perfectly reasonable response. But there are more controversies beyond this.

The digital mental wellness provider Koko, for example, used OpenAI’s technology to aid in providing responses to its platform’s users. Although co-founder Rob Morris tried to downplay ethical concerns – like whether using the tool, even in a way that was supervised by humans, violated any laws on patient safety and privacy – people reacted with anger. Morris, a former Airbnb data scientist, also pondered why we are drawn to tools like ChatGPT and why they might not yet work in peer-to-peer support settings.

Clearly, at least among tech-minded healthcare startups, there are still bugs to work out. ChatGPT has an incredibly compelling sales pitch, but it’s still only the beginning for generative AI. Ethical questions will continue to dominate, though, especially as the underlying technology evolves. Although ChatGPT was only released in November 2022, rumours suggest the substantially improved GPT-4 is right around the corner.

Featured Resources

IT best practices for accelerating the journey to carbon neutrality

Considerations and pragmatic solutions for IT executives driving sustainable IT

Free Download

The Total Economic Impact™ of IBM Spectrum Virtualize

Cost savings and business benefits enabled by storage built with IBMSpectrum Virtualize

Free download

Using application migration and modernisation to supercharge business agility and resiliency

Modernisation can propel your digital transformation to the next generation

Free Download

The strategic CFO

Why finance transformation propels business value

Free Download

Most Popular

HMRC lost nearly 50% more devices in 2022

HMRC lost nearly 50% more devices in 2022

17 Mar 2023
The big PSTN switch off: What’s happening between now and 2025?

The big PSTN switch off: What’s happening between now and 2025?

13 Mar 2023
Outlook zero day patch causes headaches for Windows admins

Outlook zero day patch causes headaches for Windows admins

15 Mar 2023