Biggest AI threat to business is to ignore it, warns industry expert

She told the roundtable discussion Human + Machine: exploring AI’s impact on business that EY had invested $1.4 billion in AI globally over three years and carried out 4,000 different AI projects.

This approach could help firms find the right AI model for the right task – and then create clear, evidence-based business cases to use AI.

The roundtable heard that large organisations see multiple applications for AI but have to balance good governance with operational risk.

Alan Nelson, Glasgow Managing Partner at CMS, said: “Everybody is talking about AI but there is a degree of nervousness and reluctance about it. We’re starting to see the power of AI and use cases emerging; businesses know what they want, but most are not there yet.”

The round table heard there could be “horror stories” around AI causing harm because some individuals and businesses don’t know how to use it responsibly – which always needs to involve managing reputation and risk.

Catriona stressed the need to “always have a human in the loop” and said the biggest skill needed to use AI effectively was critical thinking.

Learning how to prompt AI Large Language Models (LLMs) like ChatGPT effectively is also crucial – and some smaller businesses may turn away from using LLMs because they are putting poor information in and therefore getting poor results out.

Catriona explained: “Businesses need to focus on refining their inputs. When prompts to generative AI models are good and describe the context well, you should get good responses. Tell a model how to behave and use very specific inputs, so it knows what it needs to do.”

The roundtable heard that AI could pose a risk to jobs in some circumstances and must be managed well when it comes to people, with service providers needing to explain carefully the precise value of the human in the loop.

However, AI could also create more work, and more interesting work, with greater job satisfaction and productivity. It’s not yet clear what all the new AI-related jobs will look like, so there is a transition process ahead.

Some businesses have limited, or shut down employees’ access to AI, leading to an estimated 20% of staff using ‘shadow AI’ – accessing AI models without being authorised to do so.

Businesses who limit the use of AI models must be aware of the risks, said Catriona, as it could be harder to retain staff – especially younger employees who are more receptive to AI – and to increase cybersecurity risks, if more employees are using shadow AI.

Strong business leadership will be crucial. Firms were urged to put a specific person in charge of AI and told that if senior executives understand AI, they can mandate training for all staff in its effective, responsible use, and upskill and educate their entire workforce.

One approach is to create a centre of excellence within a business – including a mixture of experts in law, risk, AI and ‘creatives’ - to test and prove the value of AI.

Third party risk management is a major issue, as lots of emerging AI uses rapidly-evolving ‘black box’ technology (which produces useful information without revealing its internal workings) – and cybersecurity will become a bigger issue with ‘unknown unknowns’ ahead.

Both Catriona Campbell and Alan Nelson urged small businesses not to be overwhelmed by AI; products such as Microsoft Copilot can bring AI into a business in a straightforward, cost-effective way, while many useful free online courses and tools are available.

Above all, businesses need to be focused.  Alan added: “What are your challenges, what outcomes are you looking for and how can AI technology help?”

While Catriona concluded: “The biggest mistake is ignoring AI completely. We’re in a period of rapid change; remember that the AI we have now is the worst we will ever see in our lives.”

For more business advice go to cms.law/scotland