Creating an AI first educational Institution for Lifelong learning and re-skilling in AI

Background

I have been developing the idea of forming a new type of institution dedicated to lifelong learning and reskilling for AI. 

I am passionate about this topic of creating an AI first institution - based on my own experience of teaching AI at Oxford. 

As AI advances at an exponential rate - the trajectories of individuals, companies and countries could depend on how rapidly they adapt to the new world.   

What exactly do we mean by lifelong reskilling for AI? 

Do you teach a million people to code in Python? Or Do you teach a million people how to code Python with AI?

Focus areas

We take the later approach and we focus on the individual

We start with two focus areas

1. An open scenario driven syllabus to learn AI
2. For professionals - an institution to develop and demonstrate expertise in specific areas such as vertical LLM agents, Cognitive robotics, Autonomous AI agents and others based on the values/principles as below

Values

Here are the values that would underpin such an institution

1. Founded on community and open source - The community, alumni and open source are the foundations of the university i.e. the desire to share knowledge and to enhance learning through an open ecosystem
2. An emphasis on Lifelong learning / reskilling in AI - we do not see the institute as a two or a four year program because AI itself will be well beyond our lifetimes
3. A focus on outcomes - Sadly, education does not focus on outcomes today. An emphasis on lifelong learning of AI could change this.  
4. Hybrid learning - primarily online - but also face to face
5. The conversation is the institution - In a job, you are often focused on specific tasks, and you miss the Intellectual conversations that you have in places of learning 
6. Old fashioned inspirations emphasising teaching - A return to more traditional approaches - Pastoral care (in a secular sense) - The Oxford Tutorial system - The Indian Gurukul system etc
7. No exams .. or other standardised testing You work in a cohort covering themes. You work to collaborate with your team members to publish papers. Validated by the technical advisory board. This process has a lot of feedback and improvement. Outcome is a paper / papers - some of which you are the lead author and others you are the secondary author. 
8. Very high intellectual standards - through the advisory board and feedback 
9. Work for social - make a difference
10.  Keep a global perspective
11. A belief in Open source
12. Identify and nurture talent
13. Work with Interdisciplinary problems
14. Community and network are the lifeblood of the organization 
15. Encourage school level teachers through the open syllabus for AI 

Conclusion

Finally, I saw this interesting post in the economist via Saeed Al Dhaheri about a university called  Zhejiang University - Behind DeepSeek lies a dazzling Chinese university It models itself on Stanford, and is in the tech hotspot of Hangzhou

This shows the intellectual foundation of deepseek was a (less known) but innovative institution. This is inspiring for me

This may sound fanciful, and on overall measures of reputation and student experience by the qs World University Rankings, Zhe Da ranks 47th globally. Yet by some metrics, the university has already eclipsed many of the world’s best. It now produces more scientific papers than any other university, according to the latest Leiden ranking, a measure of the volume of research output. It is behind only Harvard in producing papers deemed to be in the top 10% of their fields.

Launch

I have shared these ideas before in posts below

We will launch this in teams.

The first stage will be the launch of the grassroots community through the Open Scenario based learning for AI - followed by the advisory board and journal. This effort will also combine my other work - vertical LLMs, maths book, open scenario based syllabus, work on creativity/autism, value of agentic workflows,  newsletter etc as videos / papers etc as a way to create the community

How to build vertical LLM agents

Setting up a journal for vertical LLM agents

Reskilling for AI using an open - scenario driven syllabus

AI research institute based on lifelong learning

How to teach AI to a ten year old with the help of chatGPT

AGI research themes I am tracking

How to teach AI using chatGPT to a fireman from Iceland

Using micro scenarios and synthetic data to teach data driven decision making  

with Dan James Ravindra Gadgil Anjali Jain Ayşe Mutlu Aditya Jaokar Samir Kulkarni