Aired (November 25, 2023): Mula sa paglikha ng pambihirang mga imahe hanggang sa kakayahan na maging parang isang tao, mapapa-bilib ka sa kontribusyon ng artificial intelligence o A.I. Pero tila kahit gaano man katayog ang potensyal ng A.I., marami pa rin ang may takot dito?
Tuklasin ang artificial intelligence sa pinakabagong dokumentaryo ni Atom Araullo.
#AIWow
#IWitness
Tuklasin ang artificial intelligence sa pinakabagong dokumentaryo ni Atom Araullo.
#AIWow
#IWitness
Category
🗞
NewsTranscript
00:00 Sometimes in our lives, there are technologies that are advancing
00:05 that are falling on our feet.
00:10 Wow! Oh my gosh! This is so cool!
00:14 Artificial Intelligence or AI is now in the mouth of many.
00:19 From computers that can talk to humans,
00:23 to apps that can create amazing images,
00:28 and AI hosts like me.
00:31 We are now surrounded by Artificial Intelligence.
00:36 What is AI? How does it work?
00:40 And how will it conquer the world?
00:55 If you are watching this documentary,
00:58 you are probably a human.
01:00 But in the not-so-distant future,
01:03 computers may be able to watch and create their own documents.
01:08 Don't you believe me?
01:11 Let's talk about one.
01:14 This is a famous AI chatbot.
01:17 We can't remember its name.
01:20 But not far away, you already know it.
01:25 If you know, you know.
01:28 Is it okay if we interview you?
01:34 The chatbot is like a person who can talk.
01:39 Can it create a documentary?
01:42 Can you make me a script for this documentary?
01:50 [Laughs]
01:52 Oh my gosh!
01:59 The chatbot immediately went out of the frame.
02:03 Oh my gosh!
02:09 It's done. It's less than 10 seconds.
02:13 Let's read it.
02:15 And wow! It has a scissor.
02:19 Wow!
02:20 This is what we want to talk about.
02:23 But don't worry.
02:26 We won't use the AI's script.
02:29 AI is already talked about on the internet.
02:35 But do ordinary Filipinos already know about it?
02:38 To predict some things like what the future will look like.
02:43 It helps people.
02:46 It's not something that is in your body.
02:49 It's artificial.
02:51 To be honest, we've been using AI for a long time.
02:55 Maybe you don't know.
02:57 It's the algorithm that runs social media networks and shopping apps.
03:03 It decides which posts you'll see on Facebook.
03:11 Or which dress or pants you'll wear when you shop online.
03:16 Here's another one.
03:19 The filter for beauty that your auntie or uncle uses.
03:24 It's also AI.
03:26 The simplest definition of artificial intelligence is any technology that mimics the intelligence of humans.
03:36 Artificial intelligence is computers mimicking human behavior.
03:41 It can speak, read, sing, decide, and drive.
03:47 Dr. Ethel Ong is a computer science expert from De La Salle University.
03:53 She has been studying AI for a long time.
03:56 So, humans imitate computers.
04:00 Is that actual intelligence?
04:03 Are machines already intelligent?
04:06 If you're looking for intelligence, you're talking to someone
04:09 and they don't understand your feelings or emotions or how it makes you feel.
04:13 But they respond.
04:15 Yes, it is intelligent.
04:17 Before we discuss how AI works,
04:22 I'll give you an additional example of how it works.
04:26 I visited the office of White Cloak Technologies, a local software development company.
04:34 They created an app called Splurge Art,
04:37 a social media platform for what they call AI art.
04:42 So what you can do is, you type words to describe what you imagine
04:48 and then the AI will create images, randomly generated images from your words.
04:53 Let's see it in action.
04:55 Let's try to generate something.
04:58 What do you want us to generate?
05:01 Ummm…
05:02 This is the hardest part.
05:06 Yeah, you need to think.
05:08 You can create anything.
05:10 We chose a theme from other users of the app to get inspiration.
05:16 Eternity together.
05:19 What do you imagine when you see that theme?
05:21 Two…
05:23 Gosh, this is…
05:26 Wow, I didn't expect this to be so hard.
05:30 I have one.
05:31 Eternity together, you type…
05:33 I need to type it in English.
05:37 Traffic in EDSA.
05:41 Traffic in EDSA.
05:43 Okay, that's Eternity.
05:46 There, it's processing.
05:51 There, it's done.
05:53 So, EDSA, Traffic in EDSA.
05:56 I think there's a boat somewhere here.
05:59 So, let's cancel the traffic.
06:01 But it's similar to EDSA.
06:03 It's confusing. There are still motorcycles here.
06:06 My gosh, it looks like Eternity Together.
06:11 I can see the images. I'm getting stressed.
06:15 Hmm, if the AI knows EDSA, does it also know who I am?
06:25 I'm…
06:26 Atom Arawlyo.
06:31 Firefighter.
06:37 Kapanipaniwala.
06:39 But it's in Culver.
06:41 What?
06:43 See?
06:44 There.
06:46 Great.
06:49 Alright.
06:51 But why is it that people don't make these images, but computers?
06:56 How?
06:58 Technology is behind the development of artificial intelligence.
07:04 But just think of it this way.
07:07 AI is designed to learn.
07:11 So before, when we do programming, normally, what we can program is what we can define step by step.
07:18 But now, it's so complicated that only humans can do.
07:21 So for certain tasks, that type of programming is not possible,
07:25 what we do instead in AI is we're programming the code on how to learn.
07:30 It's just like a human, right?
07:32 That's why it's called artificial intelligence or "gawagawang talino."
07:36 When you start to code it, it's like a baby. It doesn't know anything.
07:40 And then as you train it over experiences, over data sets, it becomes mature and smart.
07:46 It's faster than the process of a computer.
07:48 Amazing.
07:51 But how is AI taught?
07:53 So we try to think, how do we learn?
07:56 And one way humans learn is not through rules, but through examples.
08:02 Right?
08:04 Children learn to speak, not because of school.
08:06 Children learn to speak because their parents can hear them.
08:10 So when we say we're training the model, we're just giving it a lot of examples.
08:16 How many examples?
08:17 It can only contain the entire internet.
08:20 That's its library.
08:23 Yes, that's its repository. That's its intelligence.
08:26 AI can absorb more information than just one, like us.
08:33 That's why AI has the ability to create amazing things.
08:38 In the Philippines, interest in this field is also growing.
08:44 Especially among the youth.
08:46 Dr. Ong said, "The enrollment of the Computer Science Department of De La Salle University in Manila almost doubled thanks to AI."
08:59 Some students shared their research with me.
09:04 It seems to be in line with the particular needs of Filipinos.
09:12 First on the list, is Dr. Ong's protégé, Jackie.
09:16 They created V-Hope, a companion and counseling chatbot for the mental health of youth.
09:24 We have formulated V-Hope to be able to provide a chatbot for users that they can share their stories, that can recognize their emotions.
09:36 Let's try it, Jackie.
09:40 He shared his story.
09:42 He said he doesn't have friends. What do you say to him?
09:45 We can be friends. Oh my gosh.
09:47 Oh, in fairness. Do you have any plans? You can have a date.
09:54 I am already done. So, thank you for today, V-Hope.
10:02 I have to go.
10:08 Bye. So, it was nice talking with you. Remember, I am always ready to listen.
10:13 It's good that V-Hope is not clingy.
10:16 Wait, I want to talk some more. That's the advantage of AI.
10:22 Dr. Ethel wanted to use V-Hope as a guidance counselor to help the children.
10:29 When the student detects that he is feeling bad,
10:36 he needs help.
10:38 He needs to generate a response to motivate the student that it's time to seek help.
10:43 With a human being.
10:45 Yes, with a human being.
10:47 Aaron and Eric's group was able to protect senior citizens from fake news
10:53 by recommending a reading list for the trustworthiness of news sources.
10:59 So, where does AI come in?
11:03 It just looks at those past clicked articles and checks their relations.
11:07 For example, this person has been reading a lot of sports articles,
11:10 then maybe the AI should give it more sports articles, and so on and so forth.
11:14 We also added text adjustment for the articles.
11:20 So, when they read, they can adjust the body text of the article according to their preferences.
11:28 Okay, the biggest one. The one that's most compatible with me.
11:33 There.
11:38 Okay.
11:40 Joshua made a grammar checking bot in Filipino.
11:47 It's called Valarila.
11:51 Okay, let's try to run the program.
11:56 The Adobo was eaten yesterday.
11:59 The word "eaten" is incorrect.
12:02 So, I should correct that.
12:04 The Adobo was eaten yesterday.
12:06 So, now, let's say it's not yesterday, it's tomorrow.
12:10 The Adobo will be eaten tomorrow.
12:15 Very useful for Adobo eaters.
12:18 Doc, are you proud of your student?
12:23 Yes, because number one, it's very difficult to teach Filipino to the computer.
12:30 Because the language is very dynamic, the way we form sentences is also different, and our vocabulary is also different.
12:37 Even the traffic problem, some professors and students are already looking for a solution.
12:45 I'm going to show you a video that was recorded from 2018 in EDSA.
12:52 This is a camera of MMDA.
12:55 At first, the cars are detected.
12:58 As you can see, although it's a bit blurry, it says "car", "bus", "truck".
13:02 That's our classification here.
13:05 What are the benefits of this kind of system?
13:08 Actually, we built this because we know that a lot of CCTV cameras don't utilize those videos,
13:17 for enhancing security, for monitoring incidents.
13:21 So if there's a computer that's pointed at it 24/7,
13:25 those incidents are high, it can detect them immediately, and the respondents can respond easily.
13:32 So the viewer doesn't need to watch the footage.
13:36 Because the computer doesn't get tired.
13:39 Because of the great and fast development of AI, it will change the world in the coming years and decades.
13:49 Is it revolutionary?
13:51 If you believe in what AI really can do, yes.
13:55 The computer is becoming more intelligent because machines are becoming more powerful.
14:03 And the research is open.
14:05 So imagine you have people all over the world helping make it smarter
14:10 by feeding more data, by improving its algorithms.
14:16 These are just a few of the possible AI jobs in the society.
14:20 In healthcare, it's better to diagnose diseases and develop more effective medicines.
14:27 In education, personalized lessons depend on the needs of an individual.
14:35 The manufacturing process is faster and cheaper.
14:39 Agriculture is better.
14:45 Language is translated in real-time.
14:47 And many more.
14:49 But no matter how advanced AI's potential is, its dangers are as deep as it is.
14:57 Let's go back to the chatbot we talked about earlier.
15:02 I asked for help with a simple homework.
15:14 Who didn't make a book report when they were still in school?
15:18 Write a book review about the old man and the sea by Ernest Hemingway.
15:28 Wow, okay.
15:30 That's fast.
15:33 So now I made a book review about the old man and the sea by Ernest Hemingway.
15:41 And in a snap, he already made a book review.
15:43 And if I read this, he's good.
15:47 You can even change the style of his writing.
15:51 Can you please make it more humorous?
15:59 So he made a funnier version.
16:06 In fairness, I laughed.
16:09 He's really funny.
16:11 But what's not funny is how instead of reading, thinking, and writing,
16:20 students just read to their studies.
16:25 It's more of academic dishonesty.
16:30 Meaning submitting work that you said it's yours but it's not because somebody else did it.
16:36 Is that already happening in school?
16:38 It's already happening.
16:40 Even I will admit, sometimes I would not know if this work is done by a chat GPT or a student.
16:47 Look at these images.
16:52 Wow, Tom is a good student.
16:57 Look at these images. Wow, Tom Cruise, Will Smith, Donald Trump, Johnny Depp, and other celebrities
17:06 are buying and selling street food in Divisoria.
17:10 At first glance, you might think it's a fake, right?
17:16 But if you look closely, you will see that these are fake.
17:21 Does this have fingers?
17:23 Or a glass like this?
17:27 These viral images were made by AI and Business Consultant John Orania
17:32 to show the public the power of AI.
17:36 It's amazing, right?
17:39 But can you see the problem here?
17:42 Maybe the fake is actually a fake.
17:45 Misinformation.
17:49 It's getting too good, it's reaching a point where even computer science may not be able to catch them.
17:56 People no longer think, then the false information becomes the truth.
18:01 There are people in the cybersecurity community who capture the voice in our tones,
18:06 then they will use it to scam people.
18:10 KM is a cybersecurity expert and ethical hacker.
18:17 Ethical hackers defend cyberspace or the so-called network infrastructure from attacks.
18:24 He said, we need to prepare the dangerous ends of AI, especially in the field of privacy and cybersecurity.
18:34 AI can help and at the same time, it makes a cybersecurity professional or ethical hacker's job difficult.
18:43 He can help in a way because he can make our job easier.
18:47 Bad actor, they can just go to chatGPT and create malware from there.
18:53 Can I do that? Like, chatGPT, make me malware?
18:57 Yes, definitely.
18:59 Wow. Even if I don't know how to program?
19:02 Yes, you can.
19:04 You don't need to know programming for one, but you just need to have an aim, I guess.
19:12 Aside from scam and hacking, AI also has a weapon.
19:16 It will steal opportunities and people's jobs.
19:22 Hello again. Did you miss me?
19:28 Like a rising technology in Singapore,
19:33 the AI character you met earlier is an invention of AI and tech expert Agnes Woon.
19:40 She's not a host or a reporter.
19:43 She's a virtual interviewer for job applications.
19:47 What are your strengths and weaknesses? Why should we hire you?
19:52 What are your skills? Why should you be qualified for this job?
20:00 See? She's multilingual.
20:04 She is our virtual hourtel who is the recruiter that can run the recruitment process for you
20:11 to do the job more efficiently.
20:14 I won't steal your job, Atom. At least not yet. Hahaha, joke.
20:21 Awkward.
20:24 A lot of people are concerned, let's just say, about AI.
20:30 AI can steal our jobs. What do you think about that?
20:35 For me, AI will not be replacing your jobs.
20:38 It's the people who use AI that will replace your job.
20:42 Somebody still has to use AI as a tool.
20:45 My advice for people who fear that their job is being threatened by computers
20:50 is to learn new technologies, learn AI in advance.
20:55 There are also a lot of questions about the use of AI in creative endeavors, like painting.
21:01 Are these images made by computers considered art?
21:06 For Don, yes.
21:09 Definitely. Because you still need to be creative.
21:12 You need to give the computer a good direction.
21:15 So the creativity still comes from you.
21:17 It only helps you flesh out the image that you're thinking.
21:23 How much of that is still my input?
21:25 And how much of it is something that the computer is the only one who thought of?
21:30 Very, very nice question. Because it delves deeper into what the future might look like.
21:35 Basically, the director is the human behind it.
21:39 That's my vision for mainstream AI in the future.
21:42 When it comes to AI art in particular, there are concerns, especially from artists, human artists.
21:50 They say that AI is training in our art.
21:54 So how does ownership, rights, and compensation for the artist come into that?
22:02 There's actually a debate about it.
22:04 But ultimately, I think it's all about consensus.
22:08 If I don't want the computer to be trained on my work, then it should have a say.
22:15 You can opt out.
22:18 [Music]
22:22 I want to ask this AI, "Future Philippines with AI?"
22:32 With AI, okay, nice. Let's see.
22:36 [Music]
22:41 AI says that it doesn't want too many people in the future.
22:47 So it's all buildings and trees?
22:48 Buildings, nature, lights, but no people.
22:52 I'm concerned.
22:55 [Laughs]
22:57 The AI tools that we use today are still not enough for its potential in the future.
23:05 One day, we might be able to reach what we call "general AI."
23:12 An artificial intelligence that can surpass human abilities in almost everything.
23:20 How do you feel about that?
23:22 Well, I mean, hopefully I can reach it.
23:25 So you're excited?
23:27 Very excited, yeah.
23:29 There are some who are afraid.
23:31 Because there are a lot of movies where the antagonist is the computers, right?
23:35 So the science fiction writers who write about a future where AI is dominating us may happen if we allow it to happen.
23:44 Because AI is actually smarter than a lot of the population already.
23:48 Because it has all the data. It has all our data.
23:51 It is scary, but I hope my oven toaster at home won't kill me.
23:56 At least not right now, before I throw it away.
24:00 [Laughs]
24:03 But okay, let's move away from science fiction for now.
24:06 Experts say that we need to expect the trials that AI has brought today.
24:15 I think right now, the more important question is, are we using it properly?
24:21 There needs to be regulations. There needs to be guidelines for users.
24:27 So that they know what they are using and what are the possible risks in using these things.
24:35 But do you think that as AI becomes better, do we need to catch up with our regulations and laws also to protect humans who also create?
24:51 I'm all up for regulation as long as it doesn't try to limit what tech can do.
24:57 It's more about respecting the rights of humans, basically.
25:01 Do you think that the speed of AI development, can we adapt to the speed of these changes?
25:13 I can handle it if I help.
25:17 But if not, how can we use it properly?
25:21 Because you can't replace AI, you can't stop it, it's already there.
25:25 AI is not good or bad.
25:30 What are your advantages and disadvantages?
25:33 No matter how advanced it is, for now, it is just one of the tools.
25:39 And it's still in our hands, what kind of future will it face?
25:47 Good evening, I am Atom Maraulio.
25:50 I promise, this is me.
25:53 And this is the Eyewitness.
25:56 [Music]