0:00 Transcriber: Nguyệt Nguyễn Huỳnh Thu Reviewer: Phi Yen Ngo Okay. Take a look at these two sentences.
0:04 I saw the man with the telescope.
0:07 I saw the stars with the telescope.
0:10 Here's what I want you to think about.
0:12 How did your brain do that?
0:14 How did it know that you saw a man holding a telescope, yet you were using the telescope to see the stars.
0:21 You probably didn't have to think about it or pause. You just knew automatically. So when your brain reads a sentence, it doesn't process words in isolation.
0:31 It looks at the relationship between the words, and it shines a spotlight on those words that matter the most for understanding the meaning.
0:40 So here these two sentences are almost identical, but with the telescope meant something entirely different depending on whether your spotlight was shining on the word man or the word stars.
0:53 So until recently, I wasn't able to do this.
0:56 It would process words one at a time in order, like reading along a line.
1:00 But of course, this meant that it struggled to understand the meaning, especially when meaning depended on context.
1:07 So in 2017, a team of researchers at Google published a paper called Attention is All You Need, and this introduced a new way of teaching AI how to understand language.
1:19 And they call this the attention mechanism.
1:21 And what this meant is AI then learned that rather than processing words individually, it would look at the relationships between the words and shine a spotlight on the ones that matter most for understanding the meaning, just like we do.
1:35 So essentially, AI can read those two sentences and work out the difference in a similar way to how we do.
1:42 So we are currently living through the biggest technological shift in modern history. Every day, every week, there's a new AI tool that's released.
1:52 There's a new breakthroughs that we read about in the news and across social media, and there's new predictions about how much it's going to change the way that we live and the way that we work.
2:02 And this is really exciting for a lot of people, and it is changing the way that we work and live sometimes for the better.
2:10 However, a lot of people still feel completely overwhelmed by it, which is understandable, and I'm sure a lot of people can relate because there's so much information out there that it's just it feels impossible to keep up with all of it.
2:23 Yet we are constantly told, you need to keep up with AI.
2:27 But what if it's the other way around?
2:30 So I studied psychology and neurolinguistics, and when I moved into the technology space, specifically working in AI, I. What I found most fascinating was discovering the parallels that there are between the human mind and AI.
2:44 Like the one we just looked at.
2:45 So.
2:48 Why is this important to understand?
2:53 AI was designed to mimic human intelligence.
2:56 So many of the breakthroughs in artificial intelligence actually came from studying what we could say is the most sophisticated intelligence that already existed as, um.
3:10 So when I moved into studying AI and thinking about my background in psychology and neurolinguistics, I stopped and asked myself a question.
3:18 What if these two things aren't separate at all?
3:21 What if understanding the human mind is key to understanding AI?
3:25 So I started making connections, and the attention mechanism wasn't the only example that I found.
3:32 So let me take you back to the 1930s.
3:34 B.F. Skinner, a psychologist, conducted an experiment on pigeons where he discovered that behavior is shaped by rewards and consequences.
3:42 So reward something and it strengthens, ignore it or penalize it and it fades.
3:47 So you can think of this like when you're learning to cook, for example, you cook a meal for the first time. It maybe doesn't go so well.
3:54 It's too salty or it's too bland.
3:56 So next time you you adjust it the seasoning, you change the ingredients and hopefully it turns out a little bit better.
4:03 And you keep doing this until eventually you end up with the perfect dish.
4:07 So what you're doing is you're trying, you're getting feedback, you're adjusting, and then you're improving.
4:14 Now fast forward to modern AI, when ChatGPT, for example, came about and still to this day, human trainers will sit and rate the responses that it produces.
4:24 So accurate, useful responses will receive positive feedback and those inaccurate, useless responses will receive negative feedback.
4:33 The model can then go on and use that feedback to get more feedback, adjust and improve, just like we do.
4:43 And in the 1950s and 60s, Hubel and Wiesel, two neuroscientists conducted an experiment on the visual cortex of cats because they wanted to understand how the brain processes vision.
4:55 And what they found is that the brain doesn't process vision in one go as a whole. It constructs constructs it in layers.
5:03 So starting from the simple layers like edges, lines and contrast combines those into shapes and then those into objects.
5:10 So right now, when you're looking out at this room, you're not seeing it as a whole, although you don't know about it, your brain is automatically constructing it in layers.
5:19 So decades later, computer scientists asked, how could we teach AI to process vision in a similar way?
5:27 And they essentially use that as the blueprint.
5:30 And you can now see this in the mechanisms that are in the technology, such as facial recognition.
5:36 So face ID on your phone, Apple said that there's a 1 in 1,000,000 chance of someone else being able to unlock your phone.
5:43 That's because it's so specific.
5:45 Technology is so good that it will look at the you can see your edges, the lines of your face, and know that the features belong to you and only you.
5:54 And it's so specific that sometimes your own phone won't even unlock for you.
5:58 I don't know if you've experienced that.
6:00 You can also see it in medical scans.
6:02 So this means that early signs of disease can be detected that might have been missed otherwise.
6:08 And then in self-driving cars for object detection.
6:11 So for example pedestrians.
6:15 So three examples, just three three moments where scientists have studied the human mind and unknowingly shaped the future of artificial intelligence.
6:27 But it doesn't stop there.
6:29 So your brain doesn't share.
6:30 Just share a story with AI.
6:34 It shares a structure as well.
6:35 So on the left, a extremely visual representation of your brain.
6:41 This is your biological neural network.
6:43 So billions of cells, scientifically known as neurons and synapses that are connected.
6:49 and these strengthen through experience.
6:51 And then on the right, the artificial neural networks.
6:54 So billions of digital connections and these strengthen through training.
6:59 So one inspired the other.
7:02 Yet although our connections can strengthen through experience, they can also weaken through not using them.
7:09 So there's a term called cognitive offloading, which is when we consistently offload cognitive tasks to a tool, for example, AI.
7:17 So things like navigating problem solving, critical thinking.
7:21 This means that the part of our brain that's usually responsible and is used when completing tasks gets less practice.
7:28 So essentially you can weaken them connections over time.
7:32 And this is important because we've seen a lot of overreliance on AI at the moment.
7:42 So for all the similarities and the parallels that we have looked at, and there's many more where they came from.
7:49 There is, of course, still a profound difference between artificial intelligence and human intelligence.
7:56 So what is human intelligence?
7:59 Human intelligence is not just what we know or the knowledge that we've accumulated.
8:03 It's also our ability to make connections, those connections that we make through life that lead to breakthroughs and creativity and wisdom.
8:12 They don't come from nowhere.
8:13 They come from experience.
8:16 So this raises an important question what experiences should we be careful not to lose in a world that is increasingly shaped by AI?
8:26 Conversations. Conversations can expose us to new perspectives that we might not have come across otherwise.
8:34 We might not have learned otherwise.
8:36 So around 8 to 10 years ago, when I was a lot younger, I went through a stage of feeling extremely self-conscious about a particular feature of mine, and I was researching and planning how to fix it.
8:49 Really fixated on this problem that I thought I had. I had a conversation with my dad and I don’t remember the any of the conversation really, but the last two sentences he said really stuck with me.
9:00 And he’d said, you know, you don’t actually have a problem.
9:03 Like no one else sees this problem.
9:04 It's a confidence. If you were just more confident in yourself, then you'd probably forget about it.
9:10 And it stuck with me.
9:11 And after that, almost immediately, I'd had a consultation booked for this, this big problem that I had and everything I went out, acted more confident.
9:20 It worked for me, and I ended up cancelling the consultation.
9:24 I haven't even thought about this since really.
9:27 So the reason I tell you that story is because this was a new perspective that I learnt then from my dad.
9:33 And the interesting thing is it didn't just help me through that particular problem.
9:37 It's something that I come back to again and again in a lot of different situations.
9:41 So that's just an example of how conversations with people can encounter expose us to new perspectives.
9:48 So could I have asked AI for different perspectives in that situation?
9:54 Absolutely. And a lot of people can, you know, you can take your phone out immediately and speak about a problem. You get immediate answers, you can have a full on conversation.
10:03 And I probably would have received a lot of advice, some resources, endless amount of information.
10:10 But that's all it is.
10:11 It's information. And the conversation that I had with my dad wasn't just information, it was an experience.
10:18 So it was two humans speaking.
10:20 It was emotion. And like I say, that perspective that I learnt then stayed with me.
10:26 So conversations are really important.
10:29 There's a actually a popular prompt at the moment that you may, you may have even used where you can ask, let's say ChatGPT, for example, to act as a specific person.
10:38 So you can say, here's my research paper, here's my dissertation essay, whatever it might be.
10:43 Analyze it. Read it. Give me some new perspectives.
10:46 Tell me what to do better.
10:48 Where can I go? Research to improve on it. Now this is useful.
10:52 And I thought when I heard about it, that would have been great when I was doing my degree.
10:57 But then I thought about it some more and I really.
11:01 What would be a much more valuable experience would be to actually go out and speak to some psychology professors or just people working in the field, even your peers or someone who's been through it before having those conversations, because you never know where it might lead to.
11:15 This person might become your mentor or some open new doors for you.
11:20 You could end up having a, you know, a business partner from that.
11:23 It can be conversations with anyone.
11:25 So it doesn't need to be professional.
11:27 It doesn't need to be your dad.
11:29 It can be a stranger, someone you've met at an event for the first time, someone sitting next to you on a plane.
11:35 But conversations expose us to new perspectives.
11:41 A new environments expose us to new ideas and possibilities, opportunities and people that we might not have come across otherwise.
11:51 So this picture was taken at the top of Mount Batur in Bali, and Bali was the place I chose to go solo travelling for the first time.
11:58 Like a lot of people do.
11:59 And I went on this guided tour, this guided hike of Mount Batur.
12:04 So my guide was a local 17 year old girl and we spent hours.
12:08 It was just me and her.
12:10 We spent hours talking during this hike and a lot of it was.
12:13 She was currently reading the Harry Potter books and she was wanting me to, um, explain some of the English phrases that were used throughout it.
12:21 So essentially the whole three hours turned into a bit of an English lesson.
12:25 Now it was really fun.
12:26 And, but once the hike had done and the experience was over, it wasn't something I really thought about a lot until a few months later when I was at home, I came across an advert to teach English as a foreign language to become certified.
12:41 And I immediately thought back to this trip that I took this new environment that I was in and this new person that I'd met.
12:48 And that led me to actually enroll because I remembered I actually really enjoyed teaching English.
12:52 So why not? Why not take it further?
12:55 And this actually led me to then move into China as an English teacher.
12:58 Now, the reason I tell this story is because if I hadn't have gone to that new environment, I might not have well, I wouldn't, I wouldn't have come across that person.
13:07 And potentially the idea that I had and the opportunity to where it led me to.
13:11 So new environments don't need to be new countries.
13:15 It can be going to an event.
13:18 It could be working in a different coffee shop.
13:20 It could be as simple as just stepping outside of your usual routine, just doing something new and putting yourself in a new environment.
13:27 And the reason I think this is important is because it's becoming increasingly easy to have immediate access to answers.
13:34 But a lot of our most meaningful discoveries don't come from having answers immediately available to us.
13:40 They come from curiosity and exploration.
13:44 So if you're in a new place, rather than taking your phone out and researching the best places to go, immediately have a wander around and get a bit lost.
13:52 Discover things yourself.
13:54 Because in a world where AI can give us answers instantly, it's important that we don't outsource that curiosity.
14:03 And taking action. So taking action allows us to discover what we are capable of now.
14:08 AI can do a lot of things for us.
14:10 It can build businesses, teach us languages, help us to get fit.
14:14 It can write books, create presentations.
14:16 But the one thing it can't do is take action.
14:20 So my first day in China, when I arrived, I'd committed to spending a year there.
14:25 So I got off the plane, checked my phone, and my internet wasn't working.
14:28 I thought, okay, it's okay, we'll, uh, I'll go get my bags.
14:32 Waited for my bags, 45 minutes.
14:34 They never came. So I thought, okay, I'll go get some cash out of the ATM so I can at least try and get a taxi.
14:40 All the ATMs are out of order, and I couldn't find anyone who spoke English.
14:44 All the people who were on my flight had gone, so there was no one who spoke English.
14:49 Now, this was on my first day, so you can imagine how excited I was for to be spending a year there.
14:54 Now, why I tell you that is because challenges is where we discover what we are capable of.
15:00 And it can be challenges.
15:01 The daily small challenges that we come across, it can be bigger challenges. It could be sitting in a job interview or building a business, whatever it might be.
15:10 Um, AI can do a lot of things for us, but these experiences, they still belong to us.
15:16 So having the difficult conversations and taking risks and learning from failures and successes, those all still belong to us.
15:25 So not one person in this room has had the same combination of those moments.
15:31 You've all travelled different roads and made different Connections and met different people.
15:36 And you all have different perspectives on life.
15:40 So.
15:43 Use I absolutely use it.
15:45 I use it to brainstorm your ideas and elevate your critical thinking. Learn from it.
15:51 But just make sure that you you trust your own perspective and you trust perspectives of other human beings and you make your own connections.
15:59 Visit new places. Put yourself out there. Have those difficult conversations because we have to be careful not to outsource our curiosity.
16:08 So the more I learned about artificial intelligence, the more I realized how remarkable human intelligence already is.
16:16 And here's the thing.
16:17 So you spend your whole life thinking that you need to keep up with technology. But technology has spent its entire existence trying to keep up with you.
16:27 You are not trying to keep up with AI.
16:30 AI is trying to keep up with you. Thank you.
