“In the old world you might have gone hunting for information about what to teach to a Year 6 student - now you’ll be able to ask a chatbot,” says Loic Tallon, the International Baccalaureate’s chief data and digital officer.

Tallon is talking about the plans at the International Baccalaureate (IB) to let educators interrogate its Programme Resource Centre (PRC) - an online platform providing curriculum materials for IB schools - using a chatbot currently in development and being piloted.

This chatbot, he explains, will allow teachers to “ask questions of the PRC, as opposed to having to use it as a library”.

“[The PRC] is an incredible repository of learning tools and we are now placing a chatbot, an AI, on top of that to better respond to educators’ questions,” Tallon says.

Currently, he explains, the model is being tested with “a very small number of the community” to ensure “a very high bar on the integrity of the information that comes from the chatbot”.

The plan, though, is to make the chatbot available to all IB schools by the end of June.

Making best use of AI in schools

It is, arguably, an example of the IB living its values by embracing technology where it can bring benefits for its schools, teachers and learners, but with caution as the watchword.

This was something at the heart of the IB’s move last week to publish five draft design principles for AI that make it clear that while young people’s privacy and emotional, social and cognitive development must be prioritised, schools should pursue “responsible adoption” of AI. It argues that “both inaction and recklessness fail…learners”.

As such, the five principles, Tallon says, are necessary because “if we are really serious about looking after young people, we need to create these common rules”.

The rise of smartphones and social media - and the moves afoot to now put the genie back in the bottle with, for instance, the social media ban for under-16s in Australia and stricter rules around phones in schools - have provided valuable lessons, Tallon believes.

“We are realising we need to step into these spaces,” he says.

“We are more alert now to how we make sure that we help technology to be used in a teaching environment where it genuinely is beneficial to the student.”

Drafting policies together

However, developing a set of common rules that everyone can sign up to when your curricula are used by schools in 160 countries around the world is no mean feat.

“People’s culture towards AI and technology is going to vary a great deal - what some countries consider to be acceptable, others won’t.”

Tallon gives data privacy as an example: the EU and America are “both Western contexts” but have “different expectations”.

This is why the five principles are currently in draft form and will now be developed further in consultation with IB schools.

This will include insights gathered from attendees at IB conferences in Mumbai, India, and Johannesburg, South Africa. Those unable to attend can also offer their views via an online survey.

Part of this feedback is also to help establish some “red lines” about what AI should not be used for. “Sometimes the answer should be no,” says Tallon.

As an example, the IB suggested one “red line” or “non-negotiable” boundary could relate to assessment as follows:

“Our second principle states that AI should deepen learning. A related red line could be that high-stakes decisions about grading, progression and wellbeing must never be fully automated.”

Another example, says Tallon, is the question of whether all AI used in the classroom should be managed by the teacher.

He suggests, for example, that an educator could train an AI model on the impressionist art movement - the AI could even adopt the persona of a famous impressionist painter, like Renoir.

The students could interact with the AI by talking to it and learning from it, and it could then provide information to the teacher about how they performed. This would ensure that all usage was overseen and monitored by the teacher.

Conversly, the students would not be allowed to engage with AI chatbots, like ChatGPT or Gemini, where the educator is only aware of the output. It is, explains Tallon, about visibility.

Whether this will become a red line remains to be seen.

The IB seal of approval

Nonetheless, it is clear that AI is a big part of the IB’s thinking, with Tallon saying he can envisage a future where, once these principles and red lines are in place, the organisation would have the ability to identify “edtechs that are using AI in a way that aligns with International Baccalaureate Organisation values”.

Tallon says: “We know many schools are almost overwhelmed at the number of tools they have available to them.

“Why shouldn’t we step into that space and say, ‘Hey, these are the tools that align with IB’s values and we would encourage you to use’?”

This could be something that schools welcome, particularly for AI, he adds, because “it’s an area of significant concern and top of the agenda for our schools”.

What is clear, though, is that the IB will not be telling schools to shun the technology. The IB is positive about the “incredible role technology can play if used mindfully”, Tallon says, and its PRC chatbot plan is being touted as “a practical example of these principles in action”.