Artificial Intelligence (AI) is a set of computer techniques that enable machines to perform tasks that normally require human intelligence, such as learning from data and making predictions. Adoption of AI is growing, but firms—especially small and medium‑size enterprises (SMEs)—face a mix of barriers and strategic choices.
Key adoption barriers
- Cost and payoff: High AI‑digital‑human (AI‑DH) generation and production costs lower the relative payoff compared with traditional human promotion, discouraging adoption ("Higher AI-DH-related costs reduce the relative payoff of AI-DH promotion and suppress adoption").
- Data availability: Most machine‑learning models need large, clean datasets that many small firms lack ("Most applications of machine learning algorithms require huge amounts of clean and structured data, but such data is unavailable to most small companies").
- Skills and expertise: A shortage of in‑house AI expertise is a recurrent constraint ("Skills are the second organisational constraint; the shortage of in‑house expertise repeatedly appears as a barrier").
- Leadership perception: Owner‑manager attitudes strongly influence adoption; curiosity speeds uptake while scepticism blocks it ("The chief executive’s attitudes and knowledge drive adoption in small businesses; an owner who is curious and reasonably informed can push adoption through quickly, while a sceptical owner can block it just as quickly").
- Regulatory compliance: Emerging regulations such as the EU AI Act impose additional compliance costs, disproportionately affecting smaller firms ("Regulation such as the EU Artificial Intelligence Act adds compliance burdens that are especially heavy for smaller firms").
- Hidden and integration costs: Beyond licensing, firms often underestimate integration, staff‑time, and subscription expenses ("Costs have fallen too, though hidden costs such as integration, staff time and growing subscriptions are easy to underestimate").
- Environmental footprint: AI content production can consume significant energy and water, adding a sustainability dimension to adoption decisions.
Implementation strategies and trade‑offs
- Cost‑relative evaluation: Compare AI‑DH costs with real‑human promotion costs rather than evaluating AI in isolation ("Higher real-human promotion costs improve the relative attractiveness of AI‑DH promotion").
- Leverage generative AI: Use pre‑trained, general‑purpose models that require little proprietary data to lower data barriers ("Generative AI has changed the picture here, because general‑purpose models arrive already trained and can be useful with little or no proprietary data").
- Targeted use for less‑experienced staff: Deploy AI to augment newer or less‑skilled workers, where productivity gains of about 15% have been documented ("AI helps less experienced workers most, with a reported productivity gain of about 15% concentrated among newer and less skilled agents").
- Focus on narrow, verifiable problems: Apply AI to well‑defined tasks where outcomes can be measured, reducing the risk of over‑confidence and errors outside the AI's frontier ("AI can mislead when applied to problems outside its scope, leading to over‑confident errors").
- Invest in complementary assets: Expect a lag before productivity benefits appear; invest in data infrastructure, staff training, and integration capabilities early ("Implementation requires complementary investments; productivity gains are often realised only after a delay").
- Governance and monitoring: Establish ongoing evaluation of costs, consumer response, and environmental impact, as adoption metrics alone do not capture practical value ("Policy recommendations include reducing cost barriers, providing technical training, shared service platforms, and ongoing monitoring of implementation costs and environmental impact").
Practical implications for a general audience
- For SMEs, the decision to adopt AI hinges on leadership curiosity, affordable access to data, and the ability to cover hidden integration costs.
- Governments and industry bodies can accelerate adoption by subsidising skill‑building programs, offering shared AI service platforms, and simplifying regulatory compliance.
- Organizations should start with small, well‑scoped AI pilots that augment rather than replace human judgment, monitor performance, and scale only after demonstrable benefits.
- While AI can boost efficiency and broaden market reach, firms must balance these gains against potential environmental impacts and the risk of over‑reliance on opaque algorithms.
Overall, AI adoption presents both opportunities and challenges. Understanding cost dynamics, data and skill requirements, regulatory landscapes, and strategic implementation practices can help firms navigate barriers and realise the promised advantages of AI. [1] [2]