The Gigalime Blog:
where innovation and disruption meet
Good educators are mentors, not knowledge banks
Artificial intelligence can help to personalise learning. Artificial intelligence is increasingly being deployed across different sectors. In return these programmes curate our personal lives, sorting our news timelines and arranging our photos, making music recommendations, and suggesting what time we should leave to beat the traffic. Behind the scenes, this technology is increasingly taking over day to day decision-making in business; investing in stocks, replenishing supermarket shelves and even identifying criminal suspects.
What could this technology do for education? Many EdTech experts who spoke to the Atlantis Group (https://www.varkeyfoundation.org/what-we-do/atlantis-group) argued that the future of education lies in “personalised” learning programmes driven by artificial intelligence. Such technology gathers data about a learner’s performance over time and suggests different interventions to help them improve. As one developer put it: “With our system you create a tailored, unique learning experience optimised by student needs. We’re using deep neural networks that map the knowledge state and goals of a student.”
But understanding what artificial intelligence actually knows about students’ learning is complicated. And pinning down what personalisation really means in terms of education can also be difficult. When asked by the Atlantis Group, expert contributors offered a range of different definitions. One stated: “It’s around creating personalised learning paths through competencies.” Another argued that personalisation is: “a learning process that adapts to your strengths.” A third noted that the purpose of personalised learning is simply to deliver “the right lesson to the right student at the right time.” All three examples may serve as working definitions for the developers – and maybe for the people using the products. But all speak to different ideas of what personalisation is. As one Atlantis Group member reflected: “The world ‘personalisation’ was the most frequently used idea for what a system might look like in the future, but people seemed to use it to talk about individualisation, while all of us around this table [the Atlantis Group] think of learning as a social process.”
But a personalised approach does have promise. It’s a well-established principle that effective education systems should make sure that the teaching considers the level of the student. In practice, however, it is difficult to do this for each and every student in the system. In many systems just a fraction of learners are able to keep up with the curriculum, and most fall behind. Personalisation technology offers a range of powerful new diagnostic tools to help keep learners and teachers on track..
At scale, it could help ministers to understand how learners across a country are keeping up with the curricula. Such technology may be a particularly powerful approach for remedial or supplementary learning, to help the weakest and most disadvantaged learners keep up with the curriculum. But such an approach still requires the intervention of a teacher, to ensure that students are actually learning instead of just passing badly designed routine tests. Policymakers will need to ensure that such technology serves the learner as a whole, based on a comprehensive curriculum. As one Atlantis Group member argued “Personalised teaching is only good up to a point, because we don’t want students to learn [only] what they want; we want them to have a basic knowledge of language, history, et cetera. Personalised knowledge should be a way of helping to progress through the curriculum with some flexibility.”
Data-driven learning may well be the future of education. But at what cost to learners’ privacy? Our digital world is built on data. Big data has transformed the way we do business, and the ways that industry works. It’s time to do the same in education.
What should school look like in the "new normal" post pandemic world.
The phrase “the New Normal” is being used by everybody across all stratus of society and across all business sectors. The government is pushing for people to return to work and for children to get back to school in a “new normal” environment. The focus is on how to return to the way things were pre pandemic, but in a way that protects us. A lot of time, effort and money is being spent trying to persuade us that we must strive to return to the pre Covid 19 world. This is particularly true in schools were teachers and support staff have worked tirelessly over the summer to try and get children back into education in a safe way. But should we be using this as a catalyst for change rather than trying to turn the clock back.
In a white paper called Schools of the Future, Defining New Models of Education for the Fourth Industrial Revolution, published earlier they year by the World Economic Forum, they identify the need for education models to adapt so as to equip children for the future. Without the added complication of a world pandemic existing education models have not yet taken into account how very different our world looks from the wold that our current curriculums were designed for. Advancement in technology and globalisation have transformed the world of work. As the report points out “Many of today’s school children will work in new job types that do not yet exist, most of which are likely to have an increased premium on both digital and social-emotional skills. They will be introduced to wholly new business models whose workforces are much more distributed. In an increasingly interconnected world, future workers will be expected to collaborate with peers residing in various parts of the globe, understand cultural nuances and, in many cases, use digital tools to enable these new types of interactions.”
The report identifies eight critical characteristics in learning content and experiences that are needed to deliver high-quality learning in the Fourth Industrial Revolution—“Education 4.0”: They are:
1. Global citizenship skills: Include content that focuses on building awareness about the wider world, sustain- ability and playing an active role in the global community.
2. Innovation and creativity skills: Include content that fosters skills required for innovation, including complex problem-solving, analytical thinking, creativity and sys- tems analysis.
3. Technology skills: Include content that is based on developing digital skills, including programming, digital responsibility and the use of technology.
4. Interpersonal skills: Include content that focuses on interpersonal emotional intelligence, including empathy, cooperation, negotiation, leadership and social awareness.
5. Personalized and self-paced learning: Move from a system where learning is standardized, to one based on the diverse individual needs of each learner, and exible enough to enable each learner to progress at their own pace.
6. Accessible and inclusive learning: Move from a sys- tem where learning is con ned to those with access to school buildings to one in which everyone has access to learning and is therefore inclusive.
7. Problem-based and collaborative learning: Move from process-based to project- and problem-based content delivery, requiring peer collaboration and more closely mirroring the future of work.
8. Lifelong and student-driven learning: Move from a system where learning and skilling decrease over one’s lifespan to one where everyone continuously improves on existing skills and acquires new ones based on their individual needs.
We should be using the pandemic as a catalyst for change that will enable schools to develop a curriculum that is fit for the 21st century.
A full copy of the report can be found here 33wJtdO
Returning to school, will the disadvantaged suffer the most as the attainment gap continues to grow..
How bias in Algorithms can lead to unexpected and unwanted outcomes.
Are the A level results just the beginning?
With all the controversy about the A level results and the predicted furore that is going to happen over the GSCE’S, there has been much talk in the media about the algorithm used by Ofqual. However, very little has been written on why using an algorithm can cause so much outrage. Mark Hinton, CEO of Gigalime, shares some insights into why artificial intelligence does not always delivery the expected results.
Computational models of human behaviour can have multiple forms of bias. It is now understood that machine learning requires well controlled data for training and that the data is annotated in an unbiased way. If you train a neural network to assess credit applications based upon previous human assessments, you will get a machine that performs as well as the humans and also has any systemic bias of the humans in terms of age, gender, ethnicity, postcode etc. This kind of machine learning works better as a support system for the human expert rather than as a replacement. The secret is to remove systemic biases from the humans. A geo-social-economic-political problem, not a technology problem.
As for the reassignment of teacher awarded grades in the UK, it seems the ‘algorithm’ is not published, and therefore not open to peer review, a bit like the Russian vaccine. If the desire is to avoid grade inflation then presumably the target is to keep roughly the same percentage of A*, A, B, …U as in previous years. So if there were ‘too many’ A* and A grades awarded then some must be downgraded. But which ones? What kind of prejudice went into a methodology that downgraded nobody at Eton and focussed on state schools in poorer areas? Is it too much to ask for the exact process to be published for expert review? Or would that be a threat to national security?
For more information or for a free consultation please contact firstname.lastname@example.org