SAIL: AI Policy, Top Startups, ChatGPT Use Drops
Welcome to Sensemaking, AI, and Learning (SAIL) - a (usually) weekly look at AI in education and learning.
The last few months have seen a bit of a slowdown in the aggressive AI hype that has been assaulting us since last November. No core conditions, however, have changed for education: AI will kill the essay. AI produced content can't be determined like plagiarized content i.e. give up now - it will save everyone time. Learning has shifted from declarative knowledge to a focus on the development of the whole person (beingness).
The hype has quieted, but the reality hasn't changed. GRAILE hosted an event in Denver last week to address the challenges that AI presents to higher education leadership. The discussion made it quite clear that leaders are not doing the "AI thing" in house. They are relying extensively with external partners and providers to add AI expertise.
AI and Education
AI has forced a broad discussion on what's uniquely human. If information processing and knowledge acquisition and creation can be done by AI at a much faster rate and often with better results, can we expect a focus on improving the human condition, notably our dismal mental and emotional health?
A comprehensive AI policy framework. As solid a review as I've come across on a topic that is changing weekly.
AI: Technology
Top 100 AI startups (but not really all startups). A good image for future presentations. Education isn't terribly well represented (Jasper would qualify as does Ello - an AI reading app). The tools fit into AI-development tools, cross industry, and industry specific. Foundation models are big. "these startups have raised $22 billion in 223 deals since 2019. (Microsoft’s investment in OpenAI accounts for a whopping $13 billion of that total.)"
Meta releases an open LLM. "It is a huge leap forward for open-source, and a huge blow to the closed-source providers, as using this model will offer way more customizability and way lower cost for most companies."
How I became a machine learning practitioner. Interesting piece from a few years ago by the co-founder of OpenAI.
Challenges and Applications of LLMs (pdf). Excellent. Get up to speed on where we are in roughly one doc.
New AI translates 5,000-year-old cuneiform tablets instantly. For humans, translation is "a long and laborious process — one that takes years of training and dedication to learn to do well."
ChatGPT use drops. Due to summer break? Growing competitor base? Bard continues to improve. Claude now broadly available. Pi is making inroads. How many LLM conversational agents do we need?
Apple is in the LLM game, but nothing public yet.
AI: the social lens
Shopify replacing workers with AI: "aggressively embracing AI technology, using it for various purposes, from generating product descriptions to creating virtual sidekicks and developing a new help center AI agent"
AI-tocracy A long read. Not surprising, AI will give more power to those who have power. We may be entering the space where the many can no longer hold the few to account.
Regulating AI in hiring. "new legislation is being used to determine how much say AI can have in job applications." Focus is on transparency in what employers do with AI. It's only in New York.
Ethics built in centrally, rather than an afterthought. "Anthropic also believes strongly that leading on safety can’t simply be a matter of theory and white papers — it requires building advanced models on the cutting edge of deep learning."