A digital expert has warned ChatGPT carries inherent risks for business users despite representing a "big leap forward".
Neil Jackson, a senior digital consultant at accounting and business advisory firm Hurst, added the OpenAI-developed chatbot has serious limitations at its present stage.
Mr Jackson said that while the new natural language chatbot can be effectively utilised as a guide and help, business users should not "get too carried away" because the technology currently has a high margin for error.
READ MORE: Click here to sign up to the BusinessLive North West newsletter
He said: "There’s a lot of excitement and chatter about ChatGPT, and there are already many areas where it is able to assist businesses, for example with writing short pieces of code, prompting talking points on a variety of subjects and summarising lengthy articles to their key points.
"However, while it's a big leap forward in the way in which we can communicate with AI, and is a fun introduction to the application of chat-based technology, it is not developed or specialised enough yet to be a full business tool. It does have limitations and it needs carefully managing to mitigate these.
"Users should be aware that an artificial intelligence tool such as this is still only as good as the model it is built upon and the data available to it, which can be flawed. Because of this, its output should always be validated.
"There is an inherent risk with ChatGPT, especially as it is given wider access to the internet, that it may not be able to tell fact from fiction, and will accumulate more errors and biases.
"This technology still has a high margin for error, and we are still quite a way from AI taking over the world."
"However, that doesn’t mean ChatGPT isn’t worth exploring. With an engaged community looking out for its flaws and curating the data that it is processing, we can hope that the risk of misinformation and error is minimised and that it becomes a more robust and effective tool.
"In the future, ChatGPT will get better and better as its dataset is refined and expanded, and its processing power is increased."
READ MORE:
