SAIL: Dehumanizing, Robots Breaking Fingers, Mozilla
I have been hearing about the need for faster learning for about thirty years. At various times, organizational leaders/business consultations launch books on systems thinking, white water learning, the knowledge age, and the need for us to learn faster as individuals and as organisations. I wonder if "learning" is really the right term. To me, it increasingly feels like we need to make sense of the world more effectively and that we need to make meaning out of the deluge of information that crosses our screens daily. Of course that raises questions regarding how learning differs from sensemaking.
I'll offer a subtle distinction: learning is largely about acquiring information (internally or in embodied and distributed systems) to increase our future capacity - a child learns the alphabet, an adult learns about calculus. Both increase future capacity. Sensemaking, in contrast, is more concerned with understanding relationships between entities - i.e. how pieces fit even if we haven't "learned" details and intricacies of the pieces. For example, we were simultaneously "learning" about covid in early 2020 (scientists trying to understand the virus and how it was transmitted) and making sense about how to respond and what the broader implications were to the economy, teaching online, etc. Learning about covid set a foundation for making sense about what the research meant. Perhaps it could be structured as learning is the state of acquiring information and sensemaking is about connections and relationship within information and how that in turn directs our next steps.
A few articles of interest this week:
Resisting Dehumanization in the Age of “AI” (pdf). There is growing recognition about bias and ethics and fairness (the topic later in the presentation - most presentations now include some recognition of this). Overall, focuses on key concepts that relate to "human in the loop". This remains a central challenge especially where data quality raises substantial ethics questions.
A robot breaking the finger of a 7 year old opponent made rounds on social media this week. While attempted humour about the machine being a sore loser might be comical, it does raise concerns: "are we ready to deal with the moral and legal complexities raised by human-robot interaction?"
Mozilla Foundation released their annual report, citing AI's power dynamic, in a series of podcasts (the first two are available). This resource gives a more browsable format, addressing who has power, is it fair, who is accountable, etc.
A number of years ago, PwC released a report title Seizing the Prize . They appear to update sections of the website occasionally. Notably missing in their key sectors is education. AI researchers in the learning space have a marketing and image challenge to address. We are, at best, peripheral to the AI conversation. Of the reports that I've seen, this is the most aggressive in assessing economic impact of AI. McKinsey has an almost equally ambitious declaration about the impact of AI on China's economy.
Precision education continues to be a theme, and one where AI will no doubt be expected to contribute. The intent of this ER article is to highlight Open Education Analytics - a new initiative by Microsoft, drawing on the concepts of SoLAR's Open Learning Analytics position paper (mentioned in a previous SAIL email).
AI and human collaboration is often seen as a challenge of augmenting human intelligence. What happens though, when human emotions are the missing piece to increased performance? Should there be constraints around AI using emotions to interact with humans? What are the implications of an agent in a classroom mimicking emotions to nudge students toward success?
AI creativity has been an area of interest for researchers and random humans who spend too much time on social media. DALL-E 2 started the current spike in interest with its breathtaking text to image examples (see here for "last selfie on earth"). DALL-E 2 remains largely closed, but the beta is being opened to ~1million new accounts. Midjourney is accessible now, however.
AI's contribution to human knowledge creation is accelerating. Consider recent news in physics and biology. We aren't ready as a species for this level of rapid knowledge growth. We have barely started articulating interaction models (i.e. how human and artificial cognition interact and influence one another). Consider digital humans (twins or companions) - everything we know about trust, psychology, cognition, social interaction, etc. needs to be updated to reflect the possibility of a non-human cognitive agent in our systems of trust and socialization.
Impact of AI is an ongoing concern. The investments being made are enormous. The impacts are suspect. In education, we really don't have a basic "poster child of AI impact" that we can point to. We're still largely buying on hype. One organization has an interesting angle: get results in 30 days or don't pay.