artificial intelligence - in need of an ethical layer?
Download as PPTX, PDF2 likes1,303 views
This document discusses the need for an ethical framework for artificial intelligence (AI) used in education. It notes that algorithms and AI systems are developed by humans and can reflect their biases, potentially limiting opportunities for some groups. It suggests establishing an ethics commission or requiring ethics reviews of AI systems to ensure they promote values like diversity, inclusion, and student well-being rather than just replicating existing social norms. The document also questions whether establishing ethical guidelines is even possible given that AI systems are complex and outcomes are hard to predict. It asks readers to consider what an ethical approach to AI in education might include.
artificial intelligence - in need of an ethical layer?
- 1. AI in Education in need for ethics? Inge de Waard (at gmail dot com) Slideshare.net/ignatia @ignatia http://ignatiawebs.blogspot.com
- 2. Discuss and compare ideas on implementing an ethical layer within Artificial Intelligence for education.
- 3. Artificial Intelligence: mathematical models that enable communication, enhanced decision making, semantic reasoning, responding and learning between machines and humans.
- 4. Algorithms: a process or set of rules to be followed in calculations or other problem-solving operations. Algorithms are coded into software.
- 6. Algorithms are all around us. We are a product of algorithms that surround us.
- 7. Algorithms enter our homes, work, schools, institutes, habits… but in most cases they are invisible.
- 8. Being non-transparent results in unexpected outcomes: filter bubbles (un)professional hairstyle
- 10. AI risks to replicate the norm (filter bubbles prove it).Explained in part by the similar profiles of the creators of these algorithms.
- 11. Frank Pasquale (law prof) argued, “authority is increasingly expressed algorithmically.” Audrey Waters (fab thinker, talking tomorrow) wrote “Algorithms — their development and implementation — are important expressions of power and influence.”
- 12. Algocratic governance based on black boxes? Information and software systems rule.
- 13. Artificial Intelligence is a eufemistic term.
- 14. Utopian belief in AI is forgetting or neglecting human brain diversity
- 15. Building an AI that can defeat the human GO/chess champions (alphago created by Deepmind). But does it provide mental athletic well-being to the Go player? Can it read emotions?
- 16. Emotions drive learning. Affective computing is on the rise: computer science, psychology & cognitive science.
- 17. Which teacher do you remember and changed your life? Why?
- 18. What is happening with AI in education?
- 19. AI in formal education: semi-automated assessments, gamification, learning analytics, predictive analytics, scientific apps, automated student assistants, identity confirmation …
- 20. AI in informal learning: browser searches, personal apps, quantified self, learning locker based learning, course suggestions …
- 21. In short, AI in Ed can go both ways, and all the ways in between…
- 22. Positive scenario • Primary school assessment reveals a never-gonna-formally-learn student but enthusiastically yells out poems => gets a one-on-one tutor for language, and ultimately learns poetry. • Pre-school reveals personal skills compatible with satisfaction through skilled labor. Learning trajectory is provided, mentorship is arranged. • Humans are enhanced with technology => post-human is evolving and emotions are supported to lead to satisfied lives. Personal learning paths, enhancing strengths and intrinsic motivation based on enthusiasm and emotions and personal learning goals …
- 23. Negative scenario: • AI looks only for those profiles that are deemed to be able to contribute to society. The other humans are second class citizens with less opportunities. Emotions are screened for violent potential. • AI evolves and looks at humans as an inefficient species (based on existing human-build algorithms coding efficiency and moral codes such as peace must be achieved. Humans are put into reservations to protect them against themselves. AI develops into space exploring entitities.
- 24. Transparency to learn what is happening with AI in e.g. learning analytics and why => ethical rules.
- 25. What do we need to build an ethical layer?
- 26. Explained AI is a first step for understanding algorithmic effects.
- 27. Ethical layer: where do we want to go to as a society? Multiple sided stories?
- 28. We interpret the world using our moral compass: a complex set of cultural and philosophical preferences. For or against climate change and sustainable energy.
- 29. Maybe an Ethics commission at UN/Unesco. Ethics board every software output company. Ethics layer on top of AI, reviewing the AI.
- 30. Artificial Intelligence promoting inquiry, agency, activism and critical thinking & action?
- 31. Maybe it is just natural to increase the dominant norm? And does not need ethics? History has always kept mostly words from those in power. Does power always win, or do those win who we want to remember?
- 32. What would be part of your ethical layer, which outcomes of AI in education would you like to see?
Editor's Notes
- #3: A talk implies listening, and listening is linear. That is just one of the flaws of algorithms: linearity without complex dialogue
- #5: Facebook curates news based on algorithmic choices (e.g. based on likes of friends)
- #6: White young males mostly (yes, it matters). Is Artificial Intelligence a reversibly engineered brain?
- #7: Algorithms can be found in every piece of software: Turn-it-In, twitter, Big Data from & for MOOCs, browser searches, online sales… They are pervasive and are now used practically in all programmed applications.
- #8: We are in part algorithms due to the information we absorb coming frm the internet and our apps.
- #11: Easy to proof: in order to use online learning, English is used most frequently, simplified English with English jargon added. Written texts are the most frequently used form in content delivery and certainly in assessments. This leaves out non-written language based societies, and learn-by-mimicing behaviour. Do we ethically support the use of English, written assessments as proof of learning? A product is always a mirror of its creator. See community bounding: https://medium.com/@joaomilho/fixing-the-filter-bubble-e360a2c9bfdc
- #13: A black box is a system that has inputs and outputs but without the public being aware of its internal workings.
- #14: The IQ trap all over again, implying a simplistic reversed engineering option of the brain. AI is just one piece of information that fits with current technological and dominant power.
- #15: People make up humanity, through diversity No human is the same, nor their brains. Deep learning builds a “neural network”, loosely modelled on the human brain. This is composed of hundreds of thousands of neurons organised in different layers.
- #17: Only a small part of AI in education takes into account the anguish/emotion/state of being of the student or adult learner. This is called affective computing, and combines computer science, psychology and cognitive science: speech, facial expression.
- #18: Why are these teacher’s ‘good’?
- #20: Who cares about MOOC drop-outs? It just means the learner had better things to do. Focus on intrinsic interest and motivation to capture the learner, no matter where (informal and formal).
- #29: But morals are not ethics. Morality is a personal compass for right and wrong, therefor it is an internal, individual process, unlike ethics which can be expressed as a set of external rules. Ethics is about constructing or strengthening ideologies. It does not support one over the other, it just provides arguments for any ideological decision.