Don’t be so quick to let loose the cogs of AI
The AI party is in full swing, but is the business community heading for a hangover?
Unleashing AI”, the catchphrase used to launch the Government’s AI Opportunities Action Plan, doesn’t feel like an exaggeration. The global AI market is well and truly unleashed, projected to reach $1.2 trillion by the end of 2025, growing at a 26 per cent compound annual growth rate.
Here in Scotland, the AI sector is growing rapidly, with 69 per cent of Scottish tech firms identifying AI and machine learning as critical growth opportunities, up from 43 per cent in 2023. Edinburgh is the only UK city to be named an AI leader in a report from Deloitte exploring how global hubs are harnessing AI, from machine learning to robotic automation AI agents.
Key contributors include incubators like the Higgs Centre and Codebase, the Scottish Government’s Techscaler programme and research and collaboration bodies such as the Bayes Centre data science and AI hub and the Data Lab at Edinburgh University
But data on AI deployment shows that, in many cases, the numbers don’t add up. A recent global survey by Salesforce canvassed the experience of senior technology and business leaders on AI and governance. The starkest statistic is that while 67 per cent of organisations expect to maintain or increase AI spending, only 21 per cent report demonstrable benefits.
To those of us advising businesses on lawful, risk-based AI deployment, these numbers come as no surprise. They chime with TLT’s experience working with banks, corporates, tech firms and non-profits on data protection and AI. Driven by a combination of FOMO and effective PR by AI vendors, it’s easy for the lure of shiny new technology to displace less exciting sober commercial assessments of ROI.
There is a growing body of unhappy case studies of organisations engaging in sweeping redundancy programmes in favour of chatbots, only to engage in a desperate human recruitment drive after finding clients didn’t share their excitement about robot agents. Such lessons are painful, expensive and, if a business gets its ducks in a row, unnecessary.
The same survey says 43 per cent of respondents see governance, ethics and risk as their main challenges with AI. This aligns with the recent report by the FCA and Information Commissioner’s office which said data protection concerns and risks were a major blocker to AI adoption within financial services. Some 64 per cent of companies report a lack of a clear roadmap on AI implementation and adoption. Again, this is entirely consistent with our market experience.
So, what’s the solution? “AI is here to stay” is the universal mantra, with some justification. The genie is not going back in the bottle. But there is a middle way – an approach now looking more attractive than ever. Emerging best practice in this space includes robust governance, appropriate human intervention, customised AI risk assessments and, crucially, literacy programmes to ensure everyone in an organisation knows what they need to know about AI.
One influential voice advocating a balanced, considered approach is the Scottish AI Alliance, which works with businesses to facilitate adoption of AI. A key plank of their approach is to encourage organisations to start with the problem, then consider carefully whether AI is in fact the correct solution.
AI remains a huge opportunity and can and should be a net positive for all businesses. But the latest data appears to vindicate the strategic, risk-based approach advocated in Scotland. “Unleashed” is a good slogan but those who take it too literally may regret it.
Chris Elwell Sutton is a Tech, IP and Data Partner, TLT