CEOs, CFOs, need to build trust in AI at work, says insights report

Chief executive officers (CEO) and chief financial officers (CFO) need to build trust in artificial intelligence (AI) by taking steps in their organisations to manage associated risks. This becomes especially pertinent as AI plays a greater role in the accounting and financial reporting of businesses, requiring CFOs and financial controllers to be confident about the adequacy of oversight and controls of AI systems.

This is according to Association of Chartered Certified Accountants’ (ACCA) recent insights titled “AI monitor: trust” which also highlighted four socio-technical challenges of using AI.

Don’t miss: Why half of Malaysian CEOs have yet to adopt Gen AI in their work 

These include a current transparency and explainability gap towards AI adoption as the complexity of AI models can cloud decision-making. This means that professionals are less inclined to trust AI outputs if they cannot be adequately explained and justified.

As AI systems can make mistakes and show biases based on their training data, design and user demographics, AI outputs that are blindly trusted without appropriate testing and human oversight can lead to flawed decisions, added the report.

Moreover, an over-dependence on AI procedures in auditing and assurance may lead to a decline in the use of human intervention and judgement, which adds another layer of socio-technical challenges where the lines of accountability are blurred.

Interestingly, as trust is often built through human interaction and relationships, the use of AI that reduces human-to-human engagement can impact relational trust.

“As AI becomes integral to Malaysia’s business scene, it is important for business leaders to take proactive risk mitigation measures. By navigating these challenges, businesses can unlock the AI’s potential while safeguarding trust in the industry,” said Andrew Lim, portfolio head, ACCA Maritime Southeast Asia.

How can CEOs and CFOs build trust in AI?

It is important to note that the complexities and concerns in leveraging AI for businesses vary by use case. For example, when it comes to data analysis and pattern recognition, the trust issues of data bias, model interpretability and false positives or negatives can be mitigated by diverse data sets, explainable AI techniques and human review of outputs.

Similarly, when it comes to decision support and forecasting, organisations can implement model documentation, stakeholder involvement and liability frameworks to mitigate trust issues for data bias, accountability and potential for manipulation.

However, organisations should at least consider investing in AI literacy and skills development, collaborating with cross-functional teams and develop an AI governance framework to manage AI risks to foster trust.

In developing an AI governance framework, organisations are to identify specific uses of AI that support the organisation’s objectives and are consistent with its values along with considering the risks and challenges unique to one’s organisation to establish the governance mechanisms of AI.

After which, organisations must establish clear policies and oversight structures for the procurement, development and use of AI systems to form the key foundation of an AI governance framework. Data governance is particularly critical as the data used to train and operate AI models has a huge influence on their performance and fairness.

There needs to also be clear policies around AI model variation, monitoring and retaining to ensure performance remains reliable over time along with detailed audit trials and decision logs to ensure that individual AI outcomes can be examined.

Join us on 12 June 2024 for an exciting experience as Content360 makes its debut in Malaysia! Brace yourself to join the crème de la crème of the content marketing industry hailing from across the region. Immerse yourself in a dynamic atmosphere, and uncover the latest trends with thought leaders and solution providers from the realm of content.

Related articles: 
Study: Content production has increased by 56.7% with the introduction of gen AI
Meta's new gen AI feature will work for ad efficiency, but only if the starting point is right
What's keeping CEOs in Malaysia up at night?