National AI Centre: 42% of SMEs are not planning to adopt AI

There is a significant divide in artificial intelligence (AI) readiness among Australian small to medium enterprise (SME) businesses, with 42 per cent stating they are not planning to adopt AI in their business and 23 per cent are not aware of how to use the technology, leaving only 35 per cent adopting AI, according to the National AI Centre (NAIC).

The report, Responsible AI Index 2024, produced along with Fifth Quadrant, showed that AI adoption varies significantly across industries, with the health, education and manufacturing sectors seeing the highest uptake at 45 per cent. By contrast, only 6 per cent of businesses in agriculture used AI based solutions, highlighting the need for tailored, sector-specific strategies to boost AI literacy and adoption, especially in less tech-savvy industries. 

It also evaluated SMEs on the array of challenges preventing them from adopting AI that meets responsible and ethical guidelines, including skills gaps, funding issues and the rapid evolution of the technology.

NAIC’s report also looked at the critical gaps where organisations need guidance to meet AI needs, improve compliance and build trust with stakeholders.

According to the report, only 25 per cent of SMEs have engaged business leadership on responsible AI issues and just 23 per cent have implemented specific oversight and control measures.

Over half of organisations (54 per cent) that have an enterprise-wide AI strategy have their business leaders driving the strategy, compared to only a third (34 per cent) where the AI strategy is confined to specific divisions.

SMEs in the leading category of responsible AI practices are more likely to have business leaders driving AI strategy, ensuring strategic oversight and accountability.

“For these organisations, there is a greater appreciation of the competitive benefits that responsible AI practices offer, including reputation, innovation, operational efficiency and talent acquisition,” the report stated.

When it comes to safety and security, 84 per cent of organisations believe their AI systems comply with relevant privacy and security regulations. However, only 37 per cent have conducted comprehensive safety risk assessments.

While 69 per cent of organisations agree they have robust systems to minimise unfair treatment, only 35 per cent have selected fairness metrics aligned with desired outcomes.

Although most organisations believe their AI systems align with Australia’s AI Ethics Principles standards, less than a third of those surveyed have implemented the necessary practices to achieve that ambition. This gap between perception and practice is a critical finding.

In terms of explainability and contestability, 76 per cent of organisations claim they can explain how their algorithms work. However, only 39 per cent have developed supporting materials to clarify AI inputs and decision-making processes.

The mean Responsible AI Index score for Australian-based organisations is 44 out of 100, indicating significant room for improvement in the adoption of responsible AI practices.

The least implemented RAI practices primarily relate to accountability and oversight, except for the assessment of vendor claims on AI model performance.

Alarmingly, only 6 per cent of the emerging category and 12 per cent of developing category of organisations assess vendor claims on performance, suggesting a lack of diligence in verifying the accuracy and reliability of third-party AI models.

“This oversight could lead to the deployment of ineffective or biased AI systems,” the report said. “Organisations at higher maturity levels show greater appreciation of RAI’s competitive benefits, including reputation, innovation, operational efficiency and talent acquisition.”

The research also revealed a high awareness of the International Standards Organisation (ISO) AI Management System Standard and a strong intent to implement it.

This readiness lays a solid foundation for alignment with the forthcoming Australia’s Voluntary AI Safety Standard.

“While AI standards and Australia’s AI Ethics Principles are influential, small and medium businesses need more support to adopt responsible AI practices,” the report added. “Bridging these gaps will be instrumental in ensuring ethical and effective AI deployment across Australia.”