ChatGPT can reduce workload and boost productivity, but can cost a business its reputation if not used with care. While it can produce output quickly, the information is not always reliable.

It has the ability to comprehend complex questions and give relevant responses, but it is not without its quirks and limitations. There is a risk there could be factual errors in its response or that it infringes on copyright, which could lead to complaints, reputational damage or costly lawsuits.

Copyright and fact-checking

The IP Helpdesk, which supports European SMEs, highlights that AI cannot own copyright. The bot’s responses come from content previously generated by others and a ChatGPT user may unwillingly infringe on someone else’s intellectual property.

OpenAI, the Microsoft-backed creator of ChatGPT, includes a disclaimer that “ChatGPT may produce inaccurate information about people, places, or facts”.

Its helpdesk says it will occasionally make up facts or “hallucinate” outputs’ – and users can provide feedback with the ‘thumbs down’ button. Before publishing AI-generated content, it is important to check its accuracy and verify the response against credible sources.

The chatbot had until recently been limited to data available online up to March 2023, limiting its ability to give meaningful responses to any current or recent events in the news. But Open AI has announced an internet-browsing version for paying subscribers.

And it continues to develop, including with the roll-out of voice and image capabilities, allowing users to have a vocal conversation or show ChatGPT what they are talking about.

Security and SEO

OpenAI says it reviews conversations to improve its systems and for safety reasons. It may also use the content to train its models unless users opt out.

The National Cyber Security Centre advises against sharing sensitive information as the queries are stored online and face risks – such as being hacked or leaked. It is possible to buy or use AI-generated content to share online, but there are questions about its quality and effectiveness.

It may save on the time and cost of hiring a person to write, but the content may not be factually reliable or enlightening to most readers. It will also lack nuance and may not present the business in a good light. Search engines are generally sophisticated enough to detect AI content and avoid rewarding it with high visibility.

Embrace change

Businesses are using ChatGPT to boost productivity and keep costs down. It can be used to draft reports based on meeting notes, create summaries and generate marketing content.

The team at GWS Media found ChatGPT is good for sourcing initial ideas and can save time collating information. It can be a good starting point but it needs human oversight and input.

Sam Altman, OpenAI’s chief executive, has said jobs are going to “go away”. Roles which involve content creation, programming and routine and repetitive work – from writers to accountants – could be at risk. But it currently cannot read social cues, understand cultural nuance or be truly original.

In many cases, jobs are likely to be assisted and transformed, along with the creation of new roles – such as ‘AI Whisperers,’ to improve its output. AI systems have become more sophisticated and have the potential to improve efficiency and increase productivity for businesses, when used appropriately.

It would be advisable for many types of businesses to explore what value they could get from AI with an open mind. But they should not lose sight of the value of human creativity, research and analysis.