AI Ethics for Learning: Moving Beyond Principles to Practice
Perhaps the biggest roadblock organizations have (and will have) regarding AI is not the technology itself, but the governance and ethical design of programs, systems, and offerings that leverage the technology.
As we know, AI is not limited to LLMs. The field covers a host of approaches to technology including autonomy, machine learning, computer vision, smart sensors, and more. And with each application come a host of ethical challenges. Organizations who are building technologies and those who are acquiring AI technologies are now in the business of managing risk and the ethical exposure presented by poorly thought-out procurement and implementation. If your organization’s ethical policy toward AI is simply “accept the EULA”, you are already at risk.
As AI infiltrates the workforce, questions that once sounded like they came from science fiction are now questions that you are expected to be able to answer. Questions about AI recommendations, training data sets, explainability and accountability, bias, privacy, data management, and whether humans can and should trust robots.
These are not just boxes for the IT department to check off on a rubric. They are questions that deserve mindful attention. The ability to answer these questions, and more, in a way that is authoritative, transparent, and auditable will be the difference between those organizations who are able to meet the strategic demands of this new era and those that won’t.
AI Ethics Is Not About Saying "No"
It is a fallacy, and a cop-out, to say that ethics slows innovation. If anything, ethical design and an organizational strategy to support it, creates conditions in which organizations are able to understand risk and make informed business decisions before problems emerge.
Successful AI implementations have nothing to do with getting across the finish line the fastest. They are about providing clear value to stakeholders and providing transparency through meaningful governance, both technical and organizational documentation and review processes, and standards of accountability. These are not “nice to haves”. These are the value points that customers and consumers should expect (and increasingly are expecting).
We see this in legislation and regulation, such as the EU AI Act. But we also see it in the way that everyday buyers and investors are applying more scrutiny to their considerations of AI partners.
Why Learning Organizations Face Unique Challenges
Learning and training environments occupy a unique position in the AI landscape. Unlike many enterprise systems, learning technologies often collect detailed information about human behavior, performance, decision-making, competencies, readiness, and development.
While this creates great opportunities to produce more impactful learning outcomes, it also increases the standard of accountability for the providers of AI systems. If the rollout of AI continues on its current trajectory, most digital systems with which the learner engages will be pulling data and processing it in real-time to personalize learning experiences, provide on-demand mentorship, make content and activity recommendations, automate decisions and even assessments, and generate or adapt content based of learner needs.
Without clear governance and an ethical framework that stands up to scrutiny, this puts learning organizations and learning technology companies at high risk of exposure.
From Principles to Practice
Implementing an ethical framework is more than recognizing the importance of human-computer interactions, privacy and accountability, and accessibility and transparency around bias and fairness. It is about operationalizing those principles into practice.
In practical terms, this requires an accounting and inventory of AI systems and data sources as well as documentation of the decision-making and governance responsibilities necessary to assess and manage risk, monitor and message the state of systems, and demonstrate evidence that the capabilities are doing what they are supposed to be doing.
Five years ago, the IEEE Learning Technology Standards Committee started up a working group to develop a standard recommended practice relevant to AI in learning systems. After five years of research and development, the product of that working group is IEEE Std 2247.4-2025, IEEE Recommended Practice for Ethically Aligned Design of Artificial Intelligence (AI) in Adaptive Instructional Systems.
The chair of the group was Jeanine DeFalco, vice-chair of the IEEE AI Standards. The vice-chair of the working group was Yet Analytics’ CEO, Shelly Blake-Plock. Following the publication of the standard by IEEE, the pair have teamed up to provide an advisory offering that helps organizations leverage the standard’s framework approach to organizational assessment in order to move from principles to practice. The end goal is an increase in the quality of ethical design and the accountability to stakeholders during the implementation of AI in learning operations and learning technologies across organizations big or small.
Next Steps: Into the Future
Our new advisory practice offers five lanes of action, from the implementation of baseline assessment to the design and deployment of enterprise governance models.
From industry and workforce development programs, to government and non-profit deployments, to academic rollouts across higher education, we seek to guide organizations in frameworks for the design and implementation of ethical AI that are accessible, meaningful, and actionable. The result will increase value to customers, provide a measurable cultural impact, and present a clear and defensible return on investment.
Ready to discuss AI Ethics for Learning? Contact Shelly Blake-Plock directly to start the conversation.