Kabahan ka na ba sa A.I.?

Home • Forums • Dictionary • Teachers • Reader • Decks • More Tools

Click any subtitle word to view Tagalog.com dictionary results.
Computer Shortcuts: Left / Right arrows to jump 2 seconds back or forward. +Enter or Space to toggle Play/Pause button. Full Screen Mode

Paano nagagamit ang artificial intelligence para magpakalat ng chismis at disinformation? Paano rin ito magagamit para labanan ang mga ito? Join our discussion with data scientist Dominic Ligot

Christian Esguerra

Mute Like

Run time: 49:06
Has AI Subtitles

Video Transcript / Subtitles:( AI generated. About AI subtitles » )

00:00.0

Good evening guys! Happy Monday sa inyong lahat. Ako po si Christian Esguerra at welcome po sa ating episode ngayong gabi ng Facts Verse, May 8.

00:30.0

At sa start of the week, may pag-uusapan tayo ng something very interesting. Noong pinoos ko yung teaser kanina, may mga nag-message sa, Uy, iniintay namin yung topic na to, we were waiting for you to discuss this topic, may mga nag-comment, may mga nag-message, and eto para sa inyo yung topic na to kasi inisip ng iba, very technical yung artificial intelligence pero sa totoo lang nakikita natin it's all over the place so hindi natin pwedeng sabihin na this is still the future.

00:56.0

Hindi pa siyempre bakikita in the future about AI or artificial intelligence pero it's all over the place. Nakikita natin yan ngayon, yung machine learning. Dito sa ginagawa natin na podcast, dito sa ating platform na ginagamit for this live program, diba pati yung for instance yung nalalaman natin yung audience reach and demographics lahat-lahat, lahat yun related dito sa pinag-uusapan natin.

01:20.0

Pero for tonight, ang mas pag-uusapan natin, paano ba nagagamit yung artificial intelligence, yung AI na tinatawag doon sa pagpapakalat ng mga mali at malisyosong informasyon, hindi lang chismes, basically disinformation. And more importantly, pag-uusapan natin kung paano ba natin magagamit yung AI din para labanan yung disinformation.

01:42.0

So yung guest natin ngayong gabi, ito po yung nakausap natin before. I think we did already two episodes in the past regarding a related topic. Pinag-uusapan natin una yung deepfakes, yung mga muka na kinakabit sa ibang katawan tapos pulidong pulid na lately. Tapos yung sumulod na episode natin na pinag-uusapan natin yung dapat ba tayong matakot sa chat GPT.

02:04.0

So ngayon, mas pointed yung pag-uusapan natin. Ano ba yung dapat nating malaman tungkol sa connection between artificial intelligence and disinformation. So medyo nosebleed yung pag-uusapan natin. Magtanong lang kayo, just post your questions dito sa ating eksperto. I'd like to introduce to you again, to welcome to our program, si Mr. Dominic Ligot. Siya po ang data scientist.

02:32.0

I'd like to welcome again, Dominic or Doc. Thank you for joining us.

02:37.0

Good evening Christian at sa mga listeners. Nice to be back.

02:42.0

Inagahan namin ng konti kasi excited ako rito sa episode. Dapat 7.30 pa. Sinimulan natin 7.28. Yan ang pagkakaiba doc, pag sa TV to you have to wait. Pwedeng malate pero hindi pwedeng mapaaga.

02:54.0

Okay. Simulan natin sa definition of terms kasi medyo nosebleed yung pag-uusapan natin. Una-una, ano ba yung artificial intelligence? Ano ba yung dapat nating malaman dyan?

03:04.0

Well, ang pinaka basic dyan, AI na tinatawag. Ito yung paggamit ng data. Kahit anong data. Text, images. Nagsimula yan actually sa statistics.

03:17.0

And then through yung process ng tinatawag nilang machine learning, na-automate mo yung pagintindi ng mga datos na yan. And then nakakagawa ka ng forecasting dati or kunyari may picture ka ng pusa at aso, natutrain mo yung machine to detect pictures.

03:38.0

I think pinaka recent na advance ngayon sa AI na yan, yung tinatawag na generative AI. Kunyari may mga pictures ka ng mga tao dati. Nakakagawa na siya ng bagong photos.

03:48.0

Because yung deepfakes was an early example where parang nasimulate mo lang yung muka, nagpagpapalit mo yung mga muka nila. Ito, completely original images and text and sound nagagawa na niya.

04:01.0

And yung isa sa pinag-uusapan yung chat GPT, ano yan, chatbot, pwede mo nang kausapin parang tao. So products yan lahat ng AI.

04:09.0

So i-impact natin lahat yan, isa-isay natin. Yung predictive text na nakita natin in the past, kasama ba yun dito sa category na ito?

04:18.0

Or machine learning or AI? Di ba pang netetext ka, hinukulaan na yung susunod mo nasabihin.

04:23.0

Yung early version yan, yung tawag dyan GPT, in a nutshell, pinipredict niya kung ano yung next na word na lalabas. Kunyari, the cat sat on the...

04:36.0

Ano yung susunod? On the mat, on the floor, on the table. So ganoon yung early version noon.

04:40.0

Yung pinaka-latest na version yan, yung tinatawag na chat GPT, hindi lang niya pinipredict yung susunod na text.

04:49.0

Sinasagot ka niya. Basically, pwede mo syang tanongin ng kahit ano about general trivia, you can ask it for advice.

04:57.0

Kaya yung datos dyan, I think yung pinakahuli, parang trillions and trillions of data points ang ginamit nila, inabsorb nila from the internet, from text, from books.

05:07.0

Naaintindihan na niya, at least at a very simple level, kung ano yung mga sagot sa mga tanong. And I think that's where the fun begins.

05:16.0

Andyan din yung potential dangers.

05:19.0

Kaya kabahan na yung old school Google. Ito kasi yung interaction. Okay, bago tayo pumunta rin. Ito ba, ikaw muna, as a data scientist, ano ba ginagawa ng data scientist?

05:30.0

Well, truth be told, nakaka-excite siya as a person who works in data at nakakabahala. Kasi ang bottom line dyan, my opinion,

05:42.0

before 2022, bago lumabas yung ChatGPT, kailangan mo talaga ng mga experts para gumamit ng AI.

05:53.0

Yung mga tao na nag-aaral ng computer science, nag-aaral ng math. Kasi nga, komplikado gawin yan. You need programming, you need a lot of data.

06:02.0

And that's still true today. Yung gumawa nitong ChatGPT, isang data science company sa US, yung open AI.

06:10.0

Pero because naging chatbot siya, yung benefit na ng AI ngayon, hindi mo na kailangan maging super expert. As in, anyone can just go to the chatbot and ask in English.

06:22.0

Minsan nga, in other languages, pwede. May nakita akong ChatGPT in Spanish.

06:27.0

Hindi ko pa alam kung nasubukan na ng mga tao in Pilipino o Tagalog. Pero may nakita akong, parang nag-demo na sinagot siya ng ChatGPT in Becky speak. As in, parang Becky yung nagsasalita.

06:40.0

So that means, naingindihan ng chatbot. That means may example siya ng ganoong text na na-aaral dati. Ngayon, pwede na siya makipag-talaktakan in whatever slang that you want.

06:58.0

Okay. In terms muna ulit, yung AI ba yan? Interchangeable ba siya rin sa term na machine learning?

07:04.0

Well, the way I would explain it, yung machine learning, yan yung proseso. Kukuha ka ng data, aaralin niya ng algorithm basically, and gagawa siya ng prediction. Yan yung machine learning.

07:17.0

Yung AI is actually a broader, parang it's, think of it like a product. Like yung chatbot, AI na yan kasi pwede mo nang kausapin, sumasagot na siya sayo. Yung ginamit niyang proseso to get to that point was machine learning.

07:31.0

So kaya siya tinatawag na artificial intelligence, it's becoming closer to humans. Parang di nakatulad dati, kailangan mo muna magsulat ng komplikadong computer program para mapagahanan mo yung isang system. Ito literally, kinakausap mo na lang siya by writing a question.

08:02.0

Actually, siyempre, di ko alam kung ano talaga yung naging conversation niya. Actually may isa pang example niya na mayroon mayroon taong magpakamatay dahil na depressed, kausap niya yung chatbot.

08:14.0

The way I would interpret that is, sa kinaganda na ng AI ngayon, alam mo siya na chatbot siya for sure. Pero the way yung nag-interact ka sa sinasabi niya, talaga halos taon na.

08:28.0

And for all you know, artificial naman din talaga siya. As in literally, pinipredik lang niya yung pwede niyang isagot sa sinasabi mo. Pero the way it affects you, it can make you emotional, it can be thought-provoking,

08:41.0

and lalo na if you start asking it about personal things, it's possible na dahil lang doon na niwala na yung engineer. Di ko pa masabi talaga kasi ano lang naman siya, program lang naman siya.

08:52.0

Pero maybe the way na nag-interact siya doon sa program na yun, naramdaman niya na parang buhay itong kausap ko. Actually may mga companies na ganyan-ganyan, gumagawa ng mga virtual assistants.

09:03.0

Hindi yung VA na tao, as in VA na AI. Para may gusto kong kausapin. Alala mo yung unang panahon, yung Siri at si Alexa. May mga documented cases...

09:14.0

... Literally there are documented cases of people who talk to their voice assistants kasi wala silang makausap. Sumasagot naman kasi. Alexa, the weather, as in it becomes very personal. So the way it's affecting us is something interesting. I'm sure hindi yan na-predict ng mga tao.

09:39.0

So lalo yung mga tao malungkot pero they have access to this voice assistants or virtual assistants. So the way they process the information that they get from the assistants, parang tao.

09:52.0

So pwedeng ano yan? If I may use a parallelism or metaphor, parang manguhula.

10:12.0

You write yourself a letter in the future and then in the future, babasahin mo yung sinulat mo and then magre-react ka. Pero ito, you don't have to wait. You're talking to yourself basically. Nagre-react lang yung machine sa'yo.

10:24.0

I think yung pinaka for me issue dyan is never, hindi yan yung original purpose kung ba't ginawa yung mga produkto na yan. Like si Alexa, that's a e-commerce app yan. Pero somehow it's able to interact with you na parang binibigyan ka na ng counseling, binibigyan ka na ng advice.

10:44.0

And mas lalo si ChatGPT kasi kasama sa information na in-ingest dyan is mga fiction, mga novels, mga stories, mga plays. So pwede ka niyang sagutin based doon sa alam niya and that's also the danger na hindi naman yan, hindi yan kinope out ng mga tao na ginawa yan.

11:01.0

Correct me if I'm wrong, ito kasi yung basa ko dyan sa ChatGPT. Pag wala sa database niya, kung tama yung term ko, yung tinatanong mo, nasabihin niya wala siya, hindi pa niya alam yan. Kaso hindi ba siya parang self-feeding monster?

11:16.0

Kasi for instance, out of curiosity, kung narari, mga nanonood dito tatanong yan. Kung narari, I'm Dominic de la Cruz. Write a profile of Dominic de la Cruz from the Philippines, nakatira dito, ganyan-ganyan. Wala siyang data, di ba, yung ChatGPT. Pero dahil curious ka, out of curiosity, pinasok mo yung data mo, edi ibig sabihin, pasok ka na ron.

11:39.0

So there are 2 ways to appreciate that, yung tanong mo. Number 1, kung naranasan niyo ng gumamit ng ChatGPT, literally it's a chatbot, question and answer. Within yung session na yan, naaalala niya yung sinagot mo.

11:53.0

So kunyari, pwede mo siyang bigyan ng introduction. Hi, I'm Dominic, this is me, kahit hindi totoo. And then mag-interact siya as if, parang assuming that's true. Pero limited memory lang yan according to the manufacturer na pag-cancel mo na yung chat na yan, makakalimutan na niya yan supposedly. Pero technically, nakastore yan somewhere whether temporarily or permanently. So that's one side.

12:19.0

The other side naman, which I'm actually more curious about. So hindi yan, okay, the way I would treat yung mga ChatGPT, parang siyang mas advanced na data storage or database. Na dati, yung very specific information or text, kailangan mong i-query para mahuli mo yung data na yan.

12:40.0

Kunyari, what's the capital of the Philippines? Manila is the capital of the Philippines. Kasi somewhere doon sa database, makakastore yan. Pero the way itong mga chatbot na ito works, hindi niya hinuhugot yung specific information na yan specifically. Ang ginagawa niya, ina-assemble niya yung answer based doon sa whatever patterns na nastore na niya.

13:00.0

So that's why potentially, nagsisinungaling siya. Kasi yung lalabas na text looks correct and siguro basing it on statistics, it's probably true. Pero actually factually wrong. Now, pros and cons. Ang con dyan, yun nga, baka it's misinformation, which we can talk about in a bit. Pero ang pro dyan is, that means pwede kang mag-explore. It can help you. It's actually a good ideation tool kasi kahit na you've never seen something before or never experienced it before.

13:30.0

So, para siyang magandang brainstorming tool. And I think hindi pa natin masyado na ma-maximize yung potential na yan. Tsaka aminin na natin, mas maraming alam si ChatGPT kesa sayo. Kasi nakuha na niya billions and trillions of information, malamang na hindi mo pa naaaral. So magandang tool din siya para malaman yung hindi mo alam.

13:51.0

And this is the problem sa mga schools na na-encounter ko. Yung ginagamit na siya ngayon ng mga students to shortcut yung mga essay, yung mga ganyan.

14:00.0

Dok, you mentioned yung mga essay. Pwede ba natin gawin ito? Kunwari, mga estudyante, hindi masyadong sophisticated yung question na binigay ng teacher, so pwede siyang i-chat GPT. Kunwari, using a bit niya, product of ChatGPT.

14:13.0

Di ba pwede mo rin naman siyang i-input sa ChatGPT kung ikaw yung teacher para magkaroon ka na idea ko na mukhang ganun yung pagkakasulat?

14:21.0

In fact may tool na pinatawag, I think zero GPT, I have to double check. It detects if it was likely written by ChatGPT. Pero ano rin yun, for me circular argument yun.

14:38.0

The next time na gamitin mo si ChatGPT, sabihin mo, write an essay as if it was me writing it. So actually somebody already did that. Bigyan mo siya ng sample ng essay mo, gagayahin ni ChatGPT yung writing style mo. Somehow it works. So hindi siya foolproof, yun yung point ko doon. Pwede pa rin siyang maloko.

14:58.0

Sa writing class kasi, Dok, madali rin ma-detect. Kunwari, makikita mo naman yung progression o regression ng estudyante kung pangirip. Biglang genius. Uy, hindi lang na grammarly ito. Baka ano pang ginawa nito. Yung grammarly, di ba ano rin yan? AID yan, tama ba?

15:14.0

Yeah, it's a simpler version. They just look at yung mga grammar patterns. Kasi medyo pagkabisado mo naman ng English, may mga tama at mali. Si ChatGPT knows that and more kasi pwede na siya maging in a very simple way creative.

15:29.0

So for me, if I were in school, actually I do teach. Kailangan mag-isip-isip na yung mga professors ngayon na kung yung kurikula mo, some of it dates back to the 80s actually, medyo na nagigame na ng ChatGPT, I think it's time na palitan mo na yung kurikula mo. Ibahin na natin yung teaching style.

15:47.0

Parang noong 1980s gusto nila ipagbawal yung calculator. That's my favorite example. Sumupagbawal yung calculator kasi masisira daw yung math. Hindi naman nangyari. Pero obviously now, like I took accounting back in the day. I can't imagine accounting without calculators. Pero once upon a time, they wanted to ban it. So I think yun yung reaction ngayon ng mga skwelaan.

16:17.0

Q2. Definition of Terms? Ano ba yung pagkakaiba ng misinformation, disinformation, and malinformation?

16:47.0