We examine nine commonly cited AI ethical principles and makes the case that what gets labeled "ethical" in AI policy is often a reflection of political perspectives and underlying assumptions rather than genuine consensus. The conversation covers the contested nature of foundational principles like fairness, transparency, and accountability, and why each one dissolves under scrutiny into competing values rather than universal ethics. We discuss the cultural and contextual variability of AI guardrails, the dominance of risk-based ethics in education policy, and feminist Duty of Care theory as a largely overlooked alternative framework.

I am interviewed on the history, present and future of MOOCs by Saffron Mccullough for the MOOC Institute. Topics range from the origin of connectivism, neural networks and what they need to work, and the impact of AI on the future of learning. For more on the MOOC Institute, see https://www.moocinstitute.org/home

This talk connects the principles of connectivism and networked learning to the opportunities and challenges of the digital Amazon. We will explore how distributed learning practices, open resources and local networks can enhance connectivity, innovation and entrepreneurship in contexts with limited infrastructure and rich sociocultural diversity and place it in the context of educating for understanding: explanation, comprehension, and the challenge for connectivism.

This is a presentation delivered internally to project partners to provide a background and overview of learning analytics. This doesn't go deeply into any particular subject, nor does it look at recent trends, but is useful as a no-nonsense introduction to the subject.

We'll look at some of the core issues underlying the fediverse, pointing to solutions if any have been found by then: identification and identity persistence, data and scale issues, and connection and consensus.

This presentation offers a taxonomy of AI application defined along the lines of the benefits or uses we obtain from AI systems: descriptive, diagnostic, predictive, prescriptive, generative, and deontic. The purpose of creating such a taxonomy is to offer an alternative to risk-based or fear-based analyses of ethics in AI.

Federated social networks, such as Mastodon and Bluesky (and there are many more) add a venue for two essential components to AI-generated learning resources: conversation and creativity. This presentation will explore how the use of federated social networks will motivate learning in an age of automated content creation. Using as a model a personal learning environment developed by the author, it will explore what learning may look like, the role of on-demand OER, and the future role of educators.

This is a workshop that introduces new open source technology that supports 'convergent' discussions in the fediverse. The session includes segments on: the fediverse and introduction to major fediverse services such as Mastodon, Bluesky, Peertube, etc.; the dynamics on online discussion including especially the characteristics of 'divergent', 'linear' and convergent discussions; and a demonstration of convergent fediverse discussion software.

The presentation outlines the technological and business model considerations addressing the challenges faced in the early years of PLEs, the elements of the system, and broader learning ecosystem requirements that will support PLEs in general. The is contrasted with the existing LMS infrastructure in order to highlight the pedagogical and economic advantages of an open and decentralized learning infrastructure, and advice will be provided to enable institutional support for a PLE ecosystem.

What is the role of AI in education and its connection to Connectivist learning theories?

This presentation summarizes major factors - technology acceptance models, risk matrices, validity measures - used to assess a survey of instructors on their acceptance of instructional technology. Transcript.

With the recent development of broadly effective generative artificial intelligence (GAI) this has led some to suggest that open learning resources can and should be developed at the time and point of need. The GAI could address each of the issues with static online learning resources (SOLR) by addressing it based on the learning need at the time of that need, as described by the specific prompt and the context of application. This presentation will provide an overview of the technology required to support such an approach. Transcript