carebot image. source: https://images.theconversation.com/files/607376/original/file-20240716-17-6g16cj.jpg?ixlib=rb-4.1.0&rect=0%2C132%2C6301%2C3146&q=45&auto=format&w=1356&h=668&fit=crop

Talk at the EthicsLab@UCT’s Neuroscience Institute

Emanating from an early announcement about the Autonomous Decisions book, a brief encounter, access as soon as the book became available, and eager reading, Dr Heidi Matisonn invited me to give a talk at their Ethics Lab at UCT’s Neuroscience Institute, located on the Groote Schuur Hospital grounds in Cape Town, on 27 February 2025. In meeting organiser Mr Tebeli’s words, summarising my abstract:

“In this talk, Prof. Keet will explore how storytelling can make complex topics in computer ethics, such as AI ethics, safety, and responsibility more accessible and engaging. Drawing from her book, Autonomous Decisions: Computing Quandaries in Short Fiction, she will highlight key ethical challenges in computing, with a special focus on medical and health informatics issues. Expect a thought-provoking discussion on topics like buggy medical devices, obsolete implants, and ethical dilemmas in care robots.”

It was a lively talk interspersed with questions, exploratory conversations, and thoughtful views that explored a range of topics that passed the revue. It being a mixed audience with students and staff from health sciences, molecular biology, and philosophy, among others, some aspects had more resonance and familiarity than others. Among others, on the fine, but important, distinction between trust versus responsible and reliable computing (including AI), whether autonomous computing systems are really autonomous, the words we use to frame an object, and the policy vacuum being not unfamiliar in the physical and mental health settings either.

If you wonder whether any of it may be of interest to you: you can read the first story of the book for free, on Lubanzi trying to get his somewhat culturally aware carebot in Claremont to serve him more wine, for which there’s also some free self-study background information available online.

(image credits: PaO_STUDIO/Shutterstock through a The Conversation article on care robots)


Slides of the DAMA SA talk about modelling languages

Last year I was invited to give a webinar for the Southern Africa chapter of the Data Management Association (DAMA), with the brief that they wanted me to cover “everything” from my The What and How of Modelling Information and Knowledge. In less than one hour. It does not fit in an hour, though looking at the slides of “Strengths and limitations of different types of modelling languages when I uploaded them recently, I did manage to cram a good amount of material into it. They may not make a lot of sense without the explanations I added during the presentation, but I’ve shared them just in case they might.

The talk commenced touching upon mind maps and models in other disciplines (using hydrology as a new running example) mainly to illustrate their ‘nice, but…’ weaknesses. Those limitations were then solved with conceptual data models and ontologies. The talk’s, and slides’, final section introduced comparisons of types of models and the, by these days nigh on obligatory, notes on whether one could end up with biassed models (spoiler: you can).

The comparison (also blogged about on the Modeling languages blog of Jordi Cabot) is, perhaps, most interesting from the perspective of professionalism in modelling in industry: which type of model to use when, and why – among the declarative models covered, that is. Choosing the wrong type of declarative model can have dire consequence. Too simple a representation can result in missed inferences that are crucial for adequate software quality; too complex a representation can take too much time to develop for a pressing problem that can get a life-saving proverbial 80% bang for the modelling buck.


More details about these topics can be found in the modelling book (written for a broader audience), which is available from Springer and many online book sellers, and even more details on various subtopics can be found in my ontology engineering open textbook and freely accessible scientific papers. I also may be able to make time for a talk.