Hi
My name is Nick McIntosh. I am a Learning Futurist at RMIT University Vietnam. My team and I work to explore emerging edTech solutions and the latest ideas in Higher Education Learning and Teaching.
Artificial Intelligence made up a big part of our work in 2023. From research and experimentation, to developing and delivering a wide range of workshops, to giving talks across Vietnam and presenting at international Learning and Teaching festivals – AI kept us busy – and we’ve been lucky enough to get some wins in there. One thing that has been particularly successful is our RMIT AI Community of Practice. With members drawn from schools and departments from all over the wider university, the AI COP has been extremely popular across Australia, Vietnam, and beyond.
A key part of this success has been our weekly digests of emerging AI content ranging from new ideas from AI thought leaders, to relevant research and journal articles, to technical updates and practical applications of AI in HE. Feedback has been fantastic with people repeatedly highlighting the need for a newsletter of this sort for the wider HE professional community. So that’s what this is intended to be:
Interested in AI but low on time? We stay on top of this so you don’t need to.
We hope this is of interest to you. If it is, click subscribe and you can expect a weekly update every Monday from now. And if you know others who might benefit from it, please forward it on.
In the meantime though, a few things from the world of AI to help you through your Wednesday:
New York Times vs OpenAI and Microsoft 👀 🍿
This is a very interesting one: the New York Times have filed suit against both OpenAI and Microsoft for billions in damages – and they might have a strong case. Long story short, OpenAI have scraped several hundred thousand NYT articles to train their GPT models and given those articles extra “weight” (reflecting the fact that the NYT is seen as a high quality source) such that ChatGPT will repeat large chunks of NYT articles verbatim without giving due credit. This raises real questions about copyright and AI model training – not least because this “Common Crawl” training dataset has been used by a range of different LLMs/AIs. Original lawsuit here and an interesting breakdown of the suit here – this is definitely a news story to watch…
Navigating our “Second Contact” with AI – as we enter the biggest election year in history
If you’ve not seen the AI Dilemma/listened to the JRE podcast where Aza Raskin and Tristan Harris (the guys behind the Netflix documentary Social Dilemma) talk AI, they’re both absolutely fantastic and well worth a listen. This is particularly timely given 2024 is the biggest election year in world history: ~4bn people across the world are going to the polls in highly consequential elections from the US to India, from Taiwan to Europe. With the current climate and the rise of AI, the potential for dis and mis-information to do damage in one form or another has never been higher. In which case, it’s concerning that very real-seeming deepfakes from credible-seeming sources (CNN) are circulating on social media already.
AI can do independent research now?!?
So AlphaFold is an absolute miracle of modern science and AI. If you’ve not learned about it already, do yourself a favour and check this out: DeepMind’s AlphaFold AI: Doing Years of Research in Minutes! As awesome as that is, it turns out that was just the beginning – in this article from Nature, authors got GPT-4 to research, plan, and conduct experiments – including doing its own research on how to use the equipment in the lab. Read more about Autonomous chemical research with large language models here.
Tired of juggling chatbots? 🤹 ChatHub has your back
Real talk: if you use only one AI, you’re missing out. When I’m writing things, my own personal workflow tends to see me using i) Pi for brainstorming and off-the-wall thinking, ii) ChatGPT for internet-based research, crunching data, and multi-modal things, and iii) Claude for high-speed long-form reading/summaries and the actual writing itself. Sometimes though, it’s great to get them working in parallel – drawing the best ideas from each in turn. Super handy but it can be very manual and time-consuming switching between different windows/AIs. A new Github repo called ChatHub allows you to easily chat with multiple AIs simutaneously. It’s still relatively new but 9k stars on Github says it’s worth a look.
“Any sufficiently advanced technology is indistinguishable from magic” (Clarke’s Third Law)
So the above quote is one that I reflect on a bit in this whole AI space – especially when I read stories like below:
- A man whose severe spinal injury had left him paralysed for 10 years is able to walk again thanks to AI; and
- A group of researchers have used AI to decipher the language humpback whales speak to each other, and recently had a 20-minute “conversation” with a humpback whale named Twain. The implications for meaningful communication with other non-human species on Earth (and potentially beyond 👽) are profound.
Hope this has been of interest and that you’ll be joining us going forward as we explore AI and the Future of AI.
After all, we really do live in the Future. Exciting times! 🤩