Part 1: AI Apprenticeship 2024 @ DiploFoundation

This post is part of the AI Apprenticeship series:

• Part 1: AI Apprenticeship 2024 @ DiploFoundation
• Part 2: Getting introduced to the invisible apprentice – AI
• Part 2.5: AI reinforcement learning vs human governance
• Part 3: Crafting AI – Building chatbots
• Part 4: Demystifying AI  
• Part 5: Is AI really that simple?
• Part 6: What string theory reveals about AI chat models
• Part 7: ‘Interpretability: From human language to DroidSpeak’
• Part 8: ‘Maths doesn’t hallucinate: Harnessing AI for governance and diplomacy’

By Dr Anita Lamprecht, supported by DiploAI and Gemini

Week 1: The apprenticeship

Last week, the eagerly awaited, hands-on AI Apprenticeship course with Diplo began. It is an unusual step to offer an educational course as an ‘apprenticeship’. The principal lecturer, Jovan Kurbalija, explained that we will not only learn about AI in theory, but also develop our chatbots with Diplo’s own AI system. These chatbots will then serve as our workpieces. How does it work?

Imagine yourself as a kind of Geppetto, a carpenter trying to create a wooden puppet that can talk, just like the fictional character Pinocchio. Instead of wood, we use a selection of large language models to bring our chatbot to digital life. Should our Pinocchio be more creative, serious, or helpful? Will it venture into the World Wide Web to collect information, or stick to the dataset provided? Our tools to shape the chatbot’s character and scope of reach are system prompts, data, and weights. Does this mean I will finally become a prompt engineer?

Don’t call me a prompt engineer

If this were simply another course on prompting and building a chatbot, I wouldn’t participate. It would be pointless to me. Developments over the past two years has shown that even the skill of prompting might soon become unnecessary. AI systems have been gathering so much experience from the countless prompts performed worldwide that they now generate their own synthetic prompts. So, why learn a skill that might soon be obsolete?

The evolving role of humans in AI

For me, this apprenticeship is about understanding the evolving role of humans in relation to AI. It is about developing a common sense of the abilities and limitations of both the technological and human aspects within our socio-technological system. It is also a great way to develop my Diplobots and outsource the memory of the knowledge I might otherwise forget.

More than a bot: Your own AI system

The best part is learning how to become more independent from individual large tech companies. While the democratisation of AI holds immense potential, recent scandals involving data harvesting and the rise of knowledge silos have cast a shadow over its promise. DiploAI demonstrates a better way by championing a bottom-up, open-source approach that prioritises inclusivity and transparency.

By starting with smaller, carefully curated datasets and actively involving diverse stakeholders, including diplomats and linguists, DiploAI fosters a more equitable and adaptable AI development process. This approach ensures that AI solutions are not only practical but also aligned with the ethical considerations and nuanced needs of diplomacy and global governance.

In a world where AI development is often shrouded in secrecy, DiploAI’s commitment to openness and collaboration provides a refreshing alternative, demonstrating that AI can be a tool for empowerment and shared progress. 

This is just the beginning of our AI Apprenticeship. Join us next week as we dive deeper into the world of DiploAI together.

The AI Apprenticeship is one of the courses offered within the Diplo AI Campus programme.