Tom Corbett
Head of Product & Privacy Counsel at Eightfold
What Is AI Literacy?
AI literacy under the EU AI Act isn’t about turning everyone into data scientists. It’s about ensuring that those working with AI systems—whether providing (developing) or deploying (using) them—have the knowledge and skills necessary to make informed decisions and to recognize potential risks and harms.
The EU AI Act's approach is contextual and risk-based. AI literacy isn’t about knowing everything about AI. It’s about understanding the specific systems you’re working with, their applications, and their implications.
Specifically Article 4 requires:
"Providers and deployers of AI systems shall take measures to ensure, to their best extent, a sufficient level of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf, taking into account their technical knowledge, experience, education and training and the context the AI systems are to be used in, and considering the persons or groups of persons on whom the AI systems are to be used."
Let’s explore these 77 words and what they mean for you and your company.
Why Does AI Literacy Matter?
The AI literacy requirement appears in the first chapter of the EU AI Act. Article 4 also applies to all AI systems, not just high-risk AI systems under the Act. The obligations of Article 4 have some of the earliest deadlines under the law. All this demonstrates EU lawmakers' recognition of AI literacy as critical to the success of human-centered AI regulation. AI literacy is fundamental. For example, human oversight—a guiding principle of the EU AI Act—requires that people actually understand how AI systems work and their impacts when deployed. Without this understanding, automation bias could take hold and human oversight become a rubber stamp instead of a safeguard.
How to Prepare
Here’s how you and your company can build a sufficient level of AI literacy before the February 2, 2025 deadline:
- Develop Training Programs. Foster a sufficient level of organization-wide AI literacy by offering core training that’s accessible to all relevant employees, not just technical teams. From there, further educate employees based on their roles. Different roles will need different levels of understanding about AI systems. Ensure employees working on or with AI systems in their business roles have a solid understanding of the systems they use and the contexts in which they are deployed. For instance, if your company uses generative AI tools in marketing, ensure your marketing team understands how the AI system generates content and its risks and limitations, for example in the area of trademark or copyright infringement for AI-generated images or logos.
- Reach out to Vendors. Ask your AI system providers (vendors) what resources they provide to their customers to help understand their AI systems and deploy them responsibly.
- Prioritize High-Risk AI Systems. Systems deemed high-risk under the Act require closer attention to ensure a sufficient level of AI literacy. Prioritize developing a sufficient level of AI literacy for these AI systems. See Article 6 and Annex III of the Act for the types of AI systems that are high-risk under the Act.
- Document Efforts. Keep track of training completion and other AI literacy-building efforts to document compliance.
The Clock Is Ticking
Enforcement of this article of the EU AI Act may initially focus on education about AI literacy, rather than penalties, but February 2, 2025 marks a turning point in how AI is governed. The EU AI Act emphasizes that AI systems must serve humanity, not the other way around. Building AI literacy now will not only help your company meet compliance obligations, it will also position it to better exploit the productivity and other benefits of AI while managing risks appropriately.
