Ensuring responsible use of AI systems through audits

Baker Tilly continues to evolve its audit methodology while building solutions to provide assurance to AI systems.

Artificial intelligence (AI) systems are becoming increasingly essential to core business models across different sectors, including finance, healthcare, technology, and HR, and ensuring transparency, equity, integrity and reliability is paramount. AI audits emerged as an important mechanism for holding AI systems accountable, reducing risks and ensuring compliance with ethical and regulatory standards like the European Union. .

Integrating AI capabilities into the auditing process has great benefits. A recent survey by the International Association for Computer Audit Education (ICAEA) shows that 69% of participants worldwide are positive and positive about using AI for audit purposes, while 78% of participants use AI to consider audit software.

The need to audit AI systems arises from concerns related to bias, explanation, security and compliance with legal frameworks. The main reasons for auditing AI systems are:

1. : AI systems can inadvertently amplify biases present in training data, leading to unfair results. Auditing helps to detect and mitigate such biases.
2. : Many AI models, especially deep learning systems, act as “black boxes”, making it difficult to understand the decision-making process. Auditing increases transparency by evaluating how the model works.
3. : AI systems are vulnerable to hostile attacks and data addiction. Audit assesses the resilience of these models to security threats.
4. AI audits are required to comply with ethical and legal standards.
5. : Organizations implementing AI audits show commitment to responsible AI use and promote trust between users and stakeholders.

Audit AI can be implemented using a variety of approaches. Each is suitable for various aspects of AI system evaluation. The main approaches include:

1. : These include reviews of AI system data, model architecture and algorithm performance. Methods include bias detection tools, explanability techniques, and security testing.
2. : These assess the governance processes surrounding AI systems development and deployment, and ensure that best practices are followed.
3. : These analyze the actual impact of AI decisions by assessing fairness, accuracy and unintended outcomes.
4. : Independent audits conducted by external organizations increase reliability.

AI audits are important to ensure ethical, fair and responsible use of AI. While current approaches provide valuable insights, audit practices must continue to evolve to accommodate advances in AI. As AI continues to move forward, it will shape the future of financial auditing and play a central role in ensuring transparency and trust in financial reporting.

Baker Tilly continuously evolves audit methodology according to international standards, while simultaneously building solutions to provide assurance to AI systems using experienced professionals and expertise from the global Baker Tilly network.

learn more

Connect Twitter