How to Empower Education with Artificial Intelligence | Luca Longo | TEDxDublinInstituteofTechnology

Summary

The speaker discusses the potential and ethical considerations of artificial intelligence, distinguishing between 'strong' and 'narrow' AI. He highlights the current educational system's disconnect with how our brains learn and advocates for the use of narrow AI to empower education. The talk envisions a future where AI technologies like virtual reality, machine learning, and gamification can personalize learning, stimulate curiosity, and unlock the potential of every learner.

Takeaways

• 🤖 Artificial Intelligence (AI) is often associated with robots, which are intelligent agents requiring capabilities like natural language processing and machine learning.
• 🚩 The concept of 'strong AI' is distant from current achievements, which are more aligned with 'narrow AI' or 'weak AI', focusing on specific tasks.
• 🌐 There's an ethical debate on the utility and safety of AI, with concerns from figures like Stephen Hawking and Elon Musk.
• 👨‍🎓 The current educational system often fails to engage students, with many feeling disengaged and a gap between teaching methods and how the brain learns.
• 📈 The relationship between teaching and learning is not linear and can be visualized as a U-shaped curve, indicating varying levels of engagement and learning.
• 🔄 There's a need for personalization in education, moving away from a one-size-fits-all approach.
• 💡 The speaker's personal journey highlights the importance of freedom of choice and self-directed learning in education.
• 📈 The 'hype cycle' of AI in education shows a progression from innovation triggers like AI and cloud computing to technologies like virtual assistants and data mining.
• 📉 The cycle also predicts a period of disillusionment as some AI technologies fail to meet expectations in education.
• 🔮 Future advancements in AI like virtual and augmented reality, mobile learning, and learning analytics are poised to transform education by offering personalized and engaging experiences.
• 🎮 Gamification, applying game principles to education, can make learning more enjoyable and effective.

Q & A

What are the two questions posed by the speaker at the end of a previous talk?

The two questions posed by the speaker are 'Can machines think?' and 'Is it really useful to let them think?'

What is the difference between 'strong artificial intelligence' and 'narrow artificial intelligence'?

Strong artificial intelligence refers to machines that can understand, learn, and apply knowledge like humans, while narrow artificial intelligence is focused on performing specific tasks and is also known as weak artificial intelligence.

What are the key features of a robot according to the speaker?

The key features of a robot include natural language processing, machine learning, automatic reasoning, knowledge representation, planning, and perception.

What does the speaker suggest about the current state of artificial intelligence in terms of achieving strong AI?

The speaker suggests that we are far away from achieving strong artificial intelligence.

What ethical concerns are raised by notable figures like Stephen Hawking and Elon Musk regarding artificial intelligence?

Stephen Hawking suggests that artificial intelligence could be the worst thing for humanity, even worse than nuclear weapons and climate change. Elon Musk warns that with artificial intelligence, we are 'summoning the demon.'

Why did the speaker not want to wake up for secondary school?

The speaker did not want to wake up for secondary school because he was disengaged and not enjoying the learning process, which reflects the broader issue of student disengagement in the current educational system.

What is the U-shaped distribution mentioned by the speaker in relation to teaching and learning?

The U-shaped distribution refers to the relationship between teaching and learning, where some students are bored (one end of the U), others struggle with complexity (the bottom of the U), and a few enjoy the class and experience optimal learning (the other end of the U).

What does the speaker believe is missing in the current educational system?

The speaker believes that the current educational system is missing personalization and a one-size-fits-all approach is no longer effective.

What is the hype cycle the speaker refers to in relation to artificial intelligence in education?

The hype cycle is a graphical representation of the maturity and adoption of various AI technologies over time, showing how they move from innovation triggers to the plateau of productivity.

What are the technologies the speaker believes will help in the disillusionment phase of AI in education?

The speaker believes that technologies like virtual assistants, chat-bots, and automated grading will be difficult to implement in the disillusionment phase due to limitations in natural language processing.

How does the speaker envision the future of education empowered by AI?

The speaker envisions a future where AI technologies like virtual and augmented reality, mobile learning, learning analytics, and gamification will enable personalized, engaging, and effective learning experiences.

Outlines

00:00

🤖 Introduction to Artificial Intelligence

The speaker begins by referencing Alan Turing's question, 'Can machines think?', and the related concept of 'strong artificial intelligence'. They define artificial intelligence as a robot or intelligent agent with capabilities like natural language processing, machine learning, and planning. The speaker acknowledges that we are far from achieving this level of AI. They also touch on the ethical concerns surrounding AI, citing warnings from Stephen Hawking and Elon Musk. However, they introduce 'narrow artificial intelligence' or 'weak AI', which is currently achievable and focuses on narrow tasks. The speaker proposes using this type of AI to empower education, drawing on personal experiences of learning and creativity.

05:01

📚 The Need for Educational Transformation

The speaker shares their own educational journey, highlighting the disengagement many students feel in the current system. They describe the gap between teaching methods and how the brain actually learns, suggesting a U-shaped distribution where some students are bored while others struggle with complexity. The speaker argues for personalization in education, contrasting their negative secondary school experience with the freedom and choice offered in college. They introduce the concept of a 'hype cycle' for AI in education, discussing various AI technologies such as cloud computing, machine learning, data visualization, and data mining. They also mention the challenges and limitations of current AI technologies like virtual assistants and automated grading.

10:04

🚀 The Future of Education with AI

The speaker envisions a future where AI technologies like virtual and augmented reality, mobile learning, learning analytics, and gamification can transform education. They emphasize the potential for self-exploration, freedom of choice, personalization of content, and nurturing of talent through these technologies. The speaker outlines a progression from the 'slope of enlightenment' to the 'plateau of productivity' in the hype cycle, where AI technologies become practical and effective in education. They conclude by expressing a desire to return to the origins of learning with a dream of what education could be with the help of AI.

Keywords

💡Artificial Intelligence

Artificial Intelligence (AI) refers to the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions. In the video, AI is discussed in the context of 'strong' and 'narrow' AI, with the speaker emphasizing the current state of AI as narrow or weak, which is focused on specific tasks. The speaker suggests using narrow AI to empower education, indicating AI's potential to transform learning experiences.

💡Robot

A robot is an intelligent agent, often a physical machine, that can perform tasks autonomously. The script mentions robots in connection with AI, suggesting that robots embody AI capabilities such as natural language processing and machine learning. However, the speaker also notes that we are far from achieving the full set of features that would define a truly intelligent robot.

💡Natural Language Processing

Natural Language Processing (NLP) is a subfield of AI that focuses on the interaction between computers and humans through natural language. The script mentions NLP as one of the features that an intelligent agent like a robot should have, highlighting its importance in AI's ability to understand and generate human language.

💡Machine Learning

Machine Learning is a subset of AI that provides systems the ability to learn and improve from experience without being explicitly programmed. The speaker discusses machine learning as a buzzword and a key component of AI, used for inducing models from data sets and making predictions, which can inform decision-making in education.

💡Ethics of Artificial Intelligence

The Ethics of Artificial Intelligence pertains to the moral principles that should guide the development and use of AI technologies. The script references concerns from notable figures like Stephen Hawking and Elon Musk about the potential negative impacts of AI, underscoring the ethical considerations that must accompany AI advancements.

💡Narrow Artificial Intelligence

Narrow Artificial Intelligence, also known as weak AI, refers to AI systems that are designed and trained for specific tasks. The speaker uses this term to describe the current state of AI, emphasizing that while we are proficient in narrow tasks, we are not yet capable of achieving 'strong' AI that can perform any intellectual task that a human being can do.

💡Educational Disengagement

Educational Disengagement describes a state where students are not interested or involved in their learning process. The script mentions that 50% of students are disengaged in the current educational system, indicating a gap between teaching methods and how students naturally learn, which AI could help bridge.

💡Personalization

Personalization in education means tailoring the learning experience to meet individual student's needs, interests, and abilities. The speaker argues for personalization as a solution to the one-size-fits-all approach in education, suggesting that AI could enable personalized learning experiences.

💡Cloud Computing

Cloud Computing refers to the delivery of computing services, including servers, storage, databases, networking, software, and analytics, over the Internet. In the script, cloud computing is mentioned as a way to provide free textbooks and online content, potentially revolutionizing access to educational resources.

💡Data Visualization

Data Visualization is the graphical representation of information and data. The speaker mentions data visualization as a technique that clusters data, informed by data mining, which is crucial for making sense of the vast amounts of educational data that AI can process.

💡Gamification

Gamification is the application of game-design elements and principles in non-game contexts, such as education, to encourage engagement. The script discusses gamification as a technique that could be used with AI to turn educational content into game-like experiences, thereby enhancing learning through enjoyment and motivation.

Highlights

The first question posed is 'Can machines think?', which relates to the concept of 'strong artificial intelligence'.

The second question is 'Is it really useful to let machines think?', which is connected to the ethics of AI.

A robot is defined as an intelligent agent requiring features like natural language processing and machine learning.

We are currently far from achieving strong artificial intelligence.

Narrow artificial intelligence, or weak AI, is the current state of AI, focusing on narrow tasks.

The speaker suggests using narrow AI to empower education.

The speaker's childhood experiences highlight the importance of play and curiosity in learning.

The speaker's experience with an Intel 286 computer sparked a passion for coding.

Disengagement among students is a significant issue in the current educational system.

There is a gap between teaching methods and how the brain actually learns.

The U-shaped distribution of teaching and learning indicates a need for personalized education.

The speaker's experience in secondary school was less engaging compared to college.

The freedom of choice in college was beneficial for the speaker's learning.

Education needs a transformation, not just a reformation.

The hype cycle is introduced as a model to understand the development of AI technologies.

Machine learning is a subfield of AI that can induce models for prediction.

Data visualization and data mining are becoming increasingly important in AI.

The speaker predicts a period of disillusionment as we realize some AI technologies are not yet feasible for education.

Virtual and augmented reality are technologies that can enhance learning.

Mobile learning allows for personalized learning experiences.

Learning analytics and intelligent tutors can personalize content to each learner.

Gamification can make learning more enjoyable and engaging.

AI has the potential to unlock the power and potential of each learner.

The speaker concludes by emphasizing the importance of dreaming and the origins of learning.

Transcripts

Translator: Federico MINELLE Reviewer: Michele Gianella

"Can machines think?"

"Is it really useful to let them think?"

These are two questions that I posed at the end of a previous talk of mine.

The first question is related -

it has been proposed by Alan Turing in 1950,

and it is related to artificial intelligence.

This is what we call "strong artificial intelligence."

If I ask you to define what artificial intelligence is,

you probably would answer with one word: robot.

And this is strictly connected to the notion of artificial intelligence

because a robot is an intelligent agent that needs to have certain features:

natural language processing, machine learning, automatic reasoning,

knowledge representation, planning and perception.

But unfortunately, we are far away from achieving this.

The second question,

"Is it really useful to let machine think?"

is connected to the ethics of artificial intelligence.

Professor Stephen Hawking says,

"Artificial intelligence

is going to be, probably, the worst thing that happens to humanity,

even worse than nuclear weapons and climate change."

Elon Musk said,

"With artificial intelligence, we are summoning the demon."

Do not panic!

There is another definition of artificial intelligence,

which is narrow artificial intelligence,

also known as weak artificial intelligence.

This is the state of the art of artificial intelligence:

we are good at narrow tasks.

And it's with this definition of artificial intelligence

that I would like to answer a question:

how can we empower education with narrow artificial intelligence?

When I was a kid, I went to a creche, like probably most of you.

I was having fun, experiencing joy.

I was self-exploring things.

I had the chance to play with my curiosity.

That was a dream.

At the age of nine, my mom bought a computer.

It was an Intel 286 processor -

yeah, the five-inch floppy disk, an old one -

but she didn't know what a computer was.

I taught myself how to code, how to write software.

I was playing with my creativity,

nurturing my passion, developing my passion, nurturing my talent.

That was a dream.

Then I moved to secondary school.

I didn't want to wake up.

I wanted to stay in bed all the time.

So, my mom came to the room, "Luca, wake up."

"Ah, I have fever."

"You don't have fever."

So she gave me a thermometer.

She went to the kitchen, preparing breakfast.

Beside my bed, I had a little lamp.

I placed the thermometer on top of the lamp:

straight away - 38.

She came back, "Oh, you have fever."

So sometimes, I managed to stay in bed.

I was cheating.

But why? Why?

Fifty percent of students or learners are disengaged

in our current educational system.

There is a gap between the way we teach and the way we learn,

the way our brain actually learns.

Imagine an x-line: teaching,

and imagine the mental workload imposed on learners from teachers.

And then imagine a y-line: learning.

The relationship between teaching and learning

follows a U-shaped distribution.

There are some students who are bored in class.

Other students perceive instructional material adverse, complex,

so their learning is minimized.

And there is a group of students who enjoy the class

and experience optimal learning.

That means the current educational system, based on standardization,

is not really working.

One-fits-all solution doesn't work anymore.

We need to have personalization.

Then, I didn't want to continue my studies after the high school.

I went to a software house.

I was playing with what I like.

I was writing software eight hours a day.

I was enjoying, but there was something missing:

I was not learning anymore.

So, at some stage, I decided to go back to college.

Why was college good for me?

I had the freedom of choice.

So, what about education today?

Do we really need a reformation?

No, we need a transformation.

So, why don't we use what we are good at in narrow artificial intelligence

to actually empower education?

I want to answer this with a hype cycle.

On the x-line, we have the time line,

the maturity of these narrow tasks

from artificial intelligence, our technologies,

and on the y-line, we have the visibility of these technologies

and the expectation we put on them.

This is not a scientific cycle in nature.

It's my understanding, interpretation of the present

and my vision for the future.

So, here we have artificial intelligence

which is a trigger of innovation for education.

We have optimization and distributed computing.

This is what we call, nowadays, cloud computing:

free textbooks, online free content.

We can play with software online and just pay a small amount of fees.

Machine learning?

A buzzword, omnipresent.

Machine learning is a subfield of artificial intelligence.

Giving a set of data, we can induce a model,

and we can use this model for prediction and informed decision-making.

We go up to the hype, we reach the peak:

data visualization.

We have many visualization techniques online that cluster data,

and these are informed by data mining.

Data mining is basically a technique borrowed from artificial intelligence

that mines data from different sources and puts them together,

and we are becoming good at data mining.

We are at the peak of inflated expectations.

Then we go in the future, and we will enter another period:

a period of disillusionment.

Here,

we realize that things that we thought can be good for empowering education

are not really feasible.

For example, virtual assistant, chat-bots, automated grading -

these technologies are nowadays difficult to be implemented

because of the limitations of artificial intelligence

at tasks like natural language processing.

Then we will enter another period in which we are enlightened:

the slope of enlightenment.

And here, we have technologies such as virtual and augmented reality.

Virtual reality is that technology that allows us to interact

with an artificial system and artificial context, with a headset.

Augmented reality is another technology from artificial intelligence

that allows us to explore real world with augmented objects,

that can actually enhance our learning.

We go up to the curve, and we have mobile learning.

We can learn from our devices

when we want, where we want, how we want.

And we will be able to create intelligent machines

that will deliver content relevant to our devices,

to our level of understanding of a topic.

And then we will have another period of time

in which we finally focus

on the narrow task of artificial intelligence

that can actually empower education,

and we will reach a period of the plateau of productivity.

Here, we have learning analytics,

metrics related to how we behave during problem-solving,

how we perceive the underlining learning task,

how we react to this task and metrics about the context.

With these metrics, we will be able to create intelligent tutors -

so, pieces of software

that, using these metrics, can deliver relevant content to us

according to our understanding of the topic, our expertise, our level.

And finally, we will have gamification.

Gamification is a technique

that applies game principles to non-game contexts.

And we will be able to create intelligent machines, intelligent agents,

that actually get the content from different sources

and turn them into a game.

I would like you to focus on the last part of the hype cycle:

the slope of enlightenment and the plateau of productivity.

With virtual and augmented reality, we can actually self-explore things,

we can play with our curiosity, creativity.

With intelligent mobile learning, we have the freedom of choice:

we can learn what we want to learn, how, where,

and we can actually follow our passion.

With learning analytics and intelligent tutors,

we can actually personalize content to each single learner.

And eventually, with intelligent gamification,

we can nurture our talent and experience joy.

So, with artificial intelligence,

we can really unlock the power and the potential

from each single learner.

I want to go back to the origins.

I want to dream again.

Thanks.

(Applause)