At William Fry’s annual end-of-year Counsel Connect seminar, in-house counsel from Ireland’s leading companies gathered to address the significant challenges posed by the regulatory obligations under the European Union Artificial Intelligence Act's, and in particular, the forthcoming AI literacy requirements which take effect in February 2025. The seminar, titled “In the AI of the Beholder: Complying with Subjective AI Literacy Requirements under the AI Act,” was led by Rachel Hayes and Dr Barry Scannell, Technology Partners at William Fry LLP.
The session provided a roadmap for businesses to achieve compliance with the Act's AI literacy requirements, which mandate that organisations ensure their staff possess the skills and knowledge required to make informed and responsible use of AI systems. Discussions covered the practical and legal implications of these obligations, offering strategies to manage risks related to data protection, intellectual property, and AI-generated inaccuracies.
AI Literacy: A Legal and Ethical Necessity
AI literacy, as defined under the AI Act, involves equipping staff with the knowledge to understand and responsibly use AI systems. Dr Barry Scannell emphasised its importance:
“AI literacy is no longer optional. From February 2025, organisations must demonstrate a sufficient level of AI understanding within their teams, not just for compliance but to ensure ethical and effective use of this transformative technology. Failure to address this could result in significant legal and reputational consequences.”
The seminar detailed practical steps for businesses, including establishing training programmes, implementing AI governance groups, and ensuring robust review mechanisms for AI-generated outputs. Participants were urged to adopt a risk-based approach to AI governance, balancing innovation with accountability.
The Risks of Generative AI: Insights and Case Studies
The event also explored the risks posed by generative AI systems, highlighting examples such as the well-publicised case of a New York lawyer who faced professional sanctions for submitting fictitious AI-generated case law. Key risks discussed included data protection concerns, potential copyright infringements during text and data mining, and the dangers of over-reliance on AI-generated outputs, which may undermine critical legal reasoning.
Rachel Hayes addressed the issue of AI-generated inaccuracies:
“Generative AI outputs are only as reliable as the inputs and oversight to which they are subjected. Organisations must instil a culture of diligence when working with AI systems, particularly if outputs are used in high-stakes environments like legal and regulatory contexts.”
AI Act Compliance: A Multidisciplinary Approach
The seminar underscored the need for a holistic approach to compliance, combining technical expertise with legal oversight. Attendees were introduced to the requirements for managing high-risk AI systems under the AI Act, including transparency obligations, human oversight, and risk management protocols. The role of contracts in clarifying data protection obligations, intellectual property rights, liability allocation, and performance standards for AI systems was also discussed extensively.
Key takeaways included:
- Businesses need to get ready. There are 8 weeks to go until the Act's AI literacy requirements become mandatory to comply with on 2 February 2025.
- The importance of AI governance structures to monitor compliance and guide responsible use, such as an AI policy.
- Practical strategies to mitigate risks, such as reviewing AI outputs for accuracy and avoiding the input of sensitive data.
- The need for ongoing education and collaboration across teams to stay aligned with legal and ethical standards.
Shaping the Future of AI Governance
With the AI Act introducing potential fines of up to 7% of global turnover for non-compliance, the stakes for businesses could not be higher. William Fry’s Counsel Connect seminar provided actionable insights to help organisations prepare for these challenges while maximising the benefits of AI technology.
In closing, Dr Scannell remarked: “AI literacy is not just a regulatory checkbox; it’s a cornerstone of sustainable AI adoption. As the EU AI Act reshapes the regulatory landscape, organisations that embrace these requirements as an opportunity for growth will position themselves as leaders in the AI-driven economy.”
