E252 How Businesses Can Build Trustworthy and Responsible AI with Disseqt AI’s Cyril Treacy

Today’s guest is Cyril Treacy, Co-Founder & Chief Operating Officer at Disseqt AI. As AI becomes increasingly embedded in business, questions around trust, transparency and compliance are becoming more important than ever. In the today’s podcast, Cyril discusses the importance of the EU AI Act, AI governance and assurance, and how businesses can build AI systems that are both innovative and trustworthy.

Cyril discusses how businesses are grappling with the sudden emergence of the EU AI Act, which is designed to protect EU citizens from biased or unjustified decisions made by AI systems. He highlights significant updates, including the provisions for transparency that came into effect on August 2nd, requiring companies to disclose when consumers are interacting with chatbots.

However, with the timeline for high-risk use cases being pushed to December, business leaders are urged to prepare now. Cyril advises that readiness is crucial even if timelines are extended, as adapting to new regulations can be a lengthy process for enterprises, especially in regulated industries like banking.

In the world of AI, governance and assurance are frequently confused but serve distinct purposes. Governance focuses on creating policies to prevent data misuse, while assurance provides the evidence necessary to prove compliance and the reduction of bias. Cyril warns against the misconception that AI governance is a one-time checklist, asserting that it must be a continuous, evolving practice.

The episode the shifts onto the impact of AI at scale, particularly in making or supporting business decisions. Cyril illuminates the complexities of maintaining accuracy and integrity, especially in cases where AI might erroneously associate consumer data leading to biases, as exemplified by Mark’s personal anecdote of his bank loan mishap.

Cyril also emphasises the goal of AI assurance – to thoroughly test AI systems to ensure robust decision-making and prevent adverse outcomes, such as bias. This rigorous testing includes monitoring new vulnerabilities, like misleading models through poetic input, to ensure AI’s reliability.

Reflecting on the startup landscape, Cyril also highlights the necessity for agility and innovation. He details how his company Disseqt continuously adapts to the rapid technological shifts by focusing on robust testing and active guardrails that provide regulatory protection and evidence of compliance.

The conversation concludes with a crucial discussion on the importance of building AI systems that are both innovative and trustworthy. Mark and Cyril examine strategies for executives to ensure AI systems align with both business goals and regulatory requirements. Establishing assurance processes and robust testing mechanisms are essential for preventing reputational damage.

AI continues to shape the modern business environment, offering both unprecedented opportunities and complex challenges. This episode highlights that leaders must navigate the evolving landscape with an emphasis on compliance, trust and adaptability.

By understanding the nuances of AI governance and assurance, businesses can leverage technology responsibly and effectively, ensuring both operational efficiency and consumer protection. Whether you’re a company looking to innovate or maintain compliance, the insights from Cyril serve as a valuable guide in this dynamic field.