Opinion | Why companies are reluctant to jump on the ChatGPT bandwagon
- For all the productivity benefits tools like ChatGPT bring, significant issues must be addressed before companies can adopt generative AI at scale
- For example, businesses can’t have AI generating inaccurate information, making miscalculations, breaching privacy and using discriminatory language
Generative AI applications span all business domains, from customer service to operations to analytics. Some companies are even planning to reduce staff in favour of AI. According to a report by Goldman Sachs, generative AI could eliminate 18 per cent of jobs globally. ChatGPT predicts it could replace 4.8 million American jobs.
But not all companies are ready to jump on the generative AI bandwagon. Despite the obvious benefits, such as increased efficiency, productivity and innovation, there are also significant issues that need to be addressed before generative AI is adopted at scale.
This can happen because such systems rely on statistical models that can get confused by similar or outdated data. Moreover, maintaining and updating the AI foundation models that such systems use can be costly and labour-intensive for small and medium-sized enterprises. Even with regular retraining, generative AI systems cannot guarantee accuracy and consistency.
Another problem is that large language models like ChatGPT are generally bad at mathematics. This poses a serious risk to financial services, where incorrect numerical results, especially dollar amounts, are unacceptable.