AI – Magic or Science?

With such a diverse audience as our alumni, it’s always tricky deciding how much detail to put into articles. Here is some further detail if you are interested in what’s behind AI,  but by all means skip down to the actual example I’ve put in the next reply.

What actually is AI?

All forms of AI rest on two core elements

• The work of a wondrously obscure 18^th century British mathematician Thomas Bayes (1701-1761), who had a profound insight on probability theory giving us what we now call Bayesian stats. His obscurity was hard-earned by publishing exactly zero papers in his lifetime. Impressive for someone who is shaping the 21^st century.
• And secondly, AI needs an almighty amount of computing power. Huge data farms that can process in parallel the vast amounts of data required to make Mr Bayes’s sums work. Techs call these computing centres ‘neural networks’, the rest of humanity tend to struggle in disbelief, ‘How much did you say they cost?’

What are Bayesian stats – how different is it to what I did at school ?

It’s a slightly different and many people argue very intuitive way of looking at stats. As so often, it’s best to describe with an example:

Step 1  – I toss a coin twice and it’s heads both times. What are the odds when I toss the coin a third time?

Most people would say ’50:50’ as the coin has no memory of what has occurred before.

Step 2 – I continue the experiment and toss the coin a further 18 times. It’s heads each time. What are the odds when I go to toss the coin a 21^st time?

Traditionally, statisticians would say ’50:50’ because the coin still doesn’t have a memory. Quite a few people might though say ‘Hang on, something’s up here, I reckon that coin is fixed.’

And that is the application of Bayesian statistics.

With more data, an eventuality that was so unlikely after 2 coin tosses that the scenario wasn’t even worth a mention becomes the most probable explanation of the events.

Thomas Bayes wrote down a simple formula to calculate this. He seized on what might seem an obvious point – if the coin was fixed after 20 throws, it was obviously fixed after 1, it’s just we had no evidence. So there WAS a probability of the coin being fixed at the start, it’s just that it was perhaps a million to one, and wasn’t worth mentioning. With each toss of the coin though, those million to one odds drop sharply, to the point where after 20 tosses (or before) it’s actually more likely that the coin is fake than a true coin is returning all those heads.

The formula Thomas Bayes worked out to do this is in fact very simple – a lot easier than those various distribution curves you might have suffered with. You’d expect now a 15 or 16 year old to do the calculation as part of their schoolwork fairly easily.

And why does this matter? Because the formula can be put into a computer program, and then with enough data we can move away from the idea that computers are binary ‘yes/no’ to a world where they think in terms of probabilities. And that is so important, as it enables pattern recognition –  whether its written words, images or voices.

AI is essentially pattern recognition on steroids. It answers questions not by knowledge, but by finding similar patterns of words in the truly vast amount of text it has access to.

How does that tie to the types of AI?

One day, there might be one thing called AI, though if there is we’ll probably call it AGI (Artificial General Intelligence) if we get there.

Until then, most of what we call AI today falls into three broad buckets:

1) Predictive AI (the quiet kind)
This is the AI that’s been around for years inside many organisations: spotting patterns in data and making predictions. For example, forecasting demand, flagging anomalies, or helping prioritise risk. You often don’t ‘chat’ to it — it runs in the background. Pattern recognition is about probability, which is where our friend Thomas comes in.

2) Generative AI (the chatty kind — GPT and friends)
This is what people mean when they say ChatGPT. It can write, summarise, translate, draft emails, and help you think through problems. It’s powerful, but it can also sound confident while being spectacularly wrong if it’s not carefully grounded in real sources. It doesn’t actually ‘understand’ very much at all. It has though read more or less everything humans have ever written (technically that’s called a Large Language Model, or LLM), and is really good at putting together sentences that appear to make sense. It does this by spotting pattens in sentences it’s read compared to what you’ve asked … which is of course just a probability problem involving huge amounts of data. Mr Bayes and those expensive neural networks again.

3) Specialist AI (the superhuman-at-one-thing kind)
These are systems designed to excel at a specific task with clear rules and feedback — like chess. That’s why AI chess engines can be astonishingly strong, while a general chat system like ChatGPT will struggle if you ask it to play.

But while chess has a fairly narrow set of rules, literally month on month the boundaries are being pushed. Google’s DeepThink is trained to solve mathematical problems. It can now (early 2026) compete at elite IMO (International Mathematical Olympiad) gold-medal level – that’s roughly equivalent to the top 50 18-year-old maths humans on the planet. A year ago, it couldn’t compete with the top 5,000. It’s quite possible that within a year, it will out-perform all humans that have ever lived in Algebra (admittedly, the easiest of the seven major mathematical disciplines, though it didn’t feel like that at school).

Could this form of AI find a cure for cancer? Yes, and many people wish everyone would just focus on that and stop all the rest of the hype.

How do they do this? The answer now should be obvious. It’s Bayesian stats with more computer power than even Turin ever dreamed of. And yes, he did dream of AI – in many senses this belongs to him at least as much as it does to Mr Bayes.

The key takeaway: AI today isn’t one thing. It’s a family of tools. It works by a combination of the application of a clever insight into probability that identities that potential factors change their importance when you consider more data, together with a ridiculous amount of computing power.

• This reply was modified 3 months, 2 weeks ago by Site Admin.
• This reply was modified 3 months, 2 weeks ago by Site Admin.

A Call to Action: turn your experience into a living knowledge base

Many alumni carry lessons that never made it into a formal report: the workaround that kept a programme running, the early warning signal everyone missed, the staff-security choice that felt impossible at the time, the coordination habit that quietly saved weeks.

The problem is obvious: when people move on, that institutional knowledge goes with them. And in our world, that’s more than just a shame.

So the idea is to extend our AI-based research assistant Eglantyne to include a new, high-value source type: Practitioner Experience Notes, written by alumni. These won’t be news and they won’t be generic theory. They’ll be first-hand, field-smart reflections — indexed, searchable, and (when relevant) prioritised in answers. Answers that we will make available through our AI Research Assistant tool to anyone that asks; we want to share knowledge, not keep it to ourselves.

This is a genuine ‘many small contributions become something powerful’ project. If even a fraction of our alumni write a few pieces, we can build a resource that helps colleagues for years.

What we’re asking you to write

A Practitioner Experience Note is a short article based on your experience. Think: ‘Here’s what happened, here’s what we did, here’s what I wish someone had told me beforehand.’

We’re looking for notes on topics like this – not pretending this list is in any way exhaustive:

• Early warning & famine: signals, thresholds, politics of declaring, data quality, community insights vs dashboards
  Example: Early warning systems for famine in Africa (1970s–80s): what we got right, what we didn’t.
• Negotiation & access: working with authorities, non-state actors, red lines, humanitarian principles in practice
• Working with communities: feedback mechanisms, safeguarding, trust rebuilding after harm, accountability under stress
• Coordination realities: cluster meetings, duplication, info-sharing culture, how to make coordination useful rather than performative
• Staff security: decision-making under pressure; travel/compound rules; local partner risk; kidnaps/armed groups; comms blackouts
  Example: Staff security decisions during the Rwanda crisis: what actually worked.
• Supply chain & logistics: last-mile delivery, customs, cold chain, local procurement pitfalls
• Programme quality: rapid assessments, MEAL under constraints, adapting interventions mid-response
• Leadership & teams: managing mixed teams, burnout, duty of care, hard personnel calls

• Advocacy & media: speaking up safely, comms during conflict, protecting staff/partners while influencing change

If you’re not sure whether your idea ‘counts’, it probably does.

Suggested length

To make this easy to contribute to and easy to use:

• Ideal length: 800–1,500 words (this size is the ‘sweet spot’ for current AI technology behind the scenes)
• Short is fine: 400–800 words (‘one lesson, one story’)

• Longer is welcome when needed: up to 2,500 words if it genuinely benefits from depth

Some of you might already have even longer articles, or feel that the topics involved are so complex they require long articles. That’s fine.  Please discuss with me and we will figure a way.

This is not meant to be academic. Clarity beats polish.

A simple recommended structure (so it’s easy to read and search)

You can follow this template loosely if you wish, but don’t feel constrained:

1. Title (clear and specific)
2. Where/when (country/region + approximate dates or period)
3. Context in 5 lines (what was happening; what your role was)
4. The problem (what made it hard/urgent/uncertain)
5. What we did (decisions, actions, approaches)
6. What worked / what didn’t (be honest — that’s the value)
7. Lessons learned (3–7 bullet points is perfect)
8. Watch-outs (risks, unintended consequences, red flags)
9. If I had to do it again… (practical advice)

1. Keywords (10–20 terms you’d want someone to search)

If you prefer a narrative style, that’s fine too — the key is that it ends with practical takeaways.

Sensible boundaries (important)

We want this to be useful and safe so the obvious guidelines given we are sharing this openly:

• No names or identifying details of individuals at risk (staff, partners, community members)
• Avoid operational specifics that could compromise security (routes, timings, sensitive methods, etc.)
• If you’re unsure, write it as if it might be widely read and we’ll help with light redaction if needed

• You can include ‘what happened’ without revealing ‘how to replicate it’

How it will work (high level)

We’ll provide a simple submission form where you can:

• select canonical country/countries and themes
• add start/end dates (even approximate)
• add keywords

• upload your article (Word/PDF — whatever’s easiest)

On our side, the system will index these as Practitioner Experience Notes. When someone asks Eglantyne a question, these notes will be treated as high-weight sources — not because they’re ‘more true’ than everything else, but because first-hand field learning is often the missing ingredient in generic summaries.

Also: the number of Practitioner Experience Notes pulled into any one answer will be small — often 1–5, rarely more — so each note can genuinely matter.

Your article will appear in ‘public domain’ searches where appropriate as a cited source.  Because of GDPR, authorship will be defined as follows:

• If the search is being made by alumni, your name can be displayed, or you can stay anonymous if you wish.
• If the search is being made by public, authorship will be ‘Save the Children Alumni Association’. If we don’t do that and published names, we would need to go into a much more formal tier of legal compliance that we want to avoid if possible.

What to write first (if you want a prompt)

If you’d like a starting point, pick one:

• A decision I got right (and why).
• A decision I got wrong (and what I learned).
• The early warning sign we missed.
• The one coordination habit that changed everything.
• A security rule that sounded bureaucratic until it wasn’t.
• How to work effectively with [X] stakeholder in [Y] context.

• Three things I wish new arrivals knew in week one.

FAQs

• I worked for Save the Children for 10 years – how many Practitioner Notes can I write? As many as you like.
• I was working for another NGO when I learned some really important lessons. Can I write about that experience? Yes, but please use discretion and make explicitly clear this was not a Save the Children programme
• Is this field work only? I specialised in Grant funding in Finance and think I learned a lot that would be useful to pass on. Yes, we are expecting the majority of articles to be field based, but there are no restrictions at all. Please write and submit.

If you’re in: reply ‘Yes – I’ll write one/some (and a rough topic. Don’t worry if someone else is writing a similar topic; the similarities and differences will be really valuable information.)

And if you’d rather talk it through before writing: also fine.

— Pete

• This reply was modified 1 month, 3 weeks ago by Pete Smith.

Pete, this is some of the most helpful stuff on AI that I’ve read; thank you!