- 2 days ago
In this compelling video, we explore the intersection of artificial intelligence and its significant role in shaping our future.
Category
ЁЯОИ
FunTranscript
00:00The year is 2025. You turn on your TV to watch the annual Academy Awards ceremony.
00:05It's time to announce this year's winner for Best Original Screenplay.
00:09And the Oscar goes to Blue Pink AI.
00:12Ladies and gentlemen, we are witnessing history.
00:15It's the first time ever that artificial intelligence wins an Oscar.
00:19And you can bet that many more will keep winning after today, the show's presenter says.
00:24How is this possible, you wonder?
00:26Two years ago, humans were beginning to perfect AI and they're already writing movies.
00:32AI is doing a lot of things today.
00:35They'll draw pictures out of thin air.
00:37They'll make up motivational quotes.
00:40They'll write entire essays based on a prompt you give them.
00:43But making movies seems like an entirely different level for AI to achieve in just two years.
00:49According to Joe Russo, director of Marvel's Avengers Endgame, this is very likely to happen.
00:55He's on the board of several AI companies and guarantees that they are evolving fast.
01:00He says that AI have the power to completely change the way we tell stories.
01:04And how is that so?
01:07Storytelling may be as old as language itself.
01:09In every culture, across different periods, we can find evidence of storytelling.
01:14Some say humans need to tell stories to make existence seem less chaotic and out of control.
01:19The earliest evidence ever found of human storytelling dates back to 30,000 years ago,
01:25when we found drawings in a cave in France.
01:28That's what we used to do.
01:29When we didn't have pen and paper, we drew on cave walls.
01:33We illustrated facts using our hands.
01:35Then we wrote on pages.
01:37We wrote plays and asked people to interpret our characters.
01:40Today, we have enough technology to illustrate and record our stories visually.
01:45So that we'll watch it as many times as we want, whenever we feel like it.
01:50But making movies is challenging.
01:52It takes a lot of time, effort, and money.
01:54That's why some movie makers are excited about the rise of the so-called generative AI.
01:59In other words, artificial intelligence is so clever that it's capable of generating scenes,
02:04scripts, characters, dialogue, soundtrack, and the list goes on and on.
02:09Picture this.
02:09You're a Hollywood producer.
02:11You receive an infinite amount of scripts every day.
02:14Some of them you know are great right away,
02:16but others you'd like to know if they're worth producing.
02:19Today, you can ask for an AI to analyze the possible success rate for that movie,
02:24how much it would make at the box office,
02:26and if it's better to do a wide or limited release.
02:29Of course, they can get it wrong.
02:32The point is, what humans take a long time to do,
02:36AIs perform in a matter of seconds.
02:38If you're an independent movie director and you need to go location scouting,
02:42you can ask an AI to do that for you.
02:45It's not going to literally go scouting,
02:47but based on what data you fed it,
02:49it will give you a list of possible locations.
02:51But perhaps the most mind-boggling feature
02:53is that an AI can write, direct, and animate an entire movie by itself.
02:58Some creatives have been taking AI for a test drive,
03:01but they still haven't given them full autonomy.
03:03Director Chad Nelson made an animated short film that got the internet talking.
03:08He asked Dali, an artificial intelligence,
03:11to generate thousands of pictures of a cute, fuzzy monster
03:14with small horns over a mossy hill forest.
03:17He experimented with prompts until he got the result he wanted to finish the short film,
03:22Critters.
03:23AI can also do things humans can't,
03:25like de-age the face of an actor on screen
03:28and make it seem as real as life itself.
03:31That's the case with Harrison Ford in the new Indiana Jones film.
03:35He appears as his younger self,
03:37even though he's 80 years old today.
03:39For OpenAI's president, Greg Brockman,
03:42AI can enhance people's movie experience
03:44by making it more immersive and interactive.
03:47Let's say you had a tough day at work.
03:49You get home and you're in the mood for watching a light-hearted rom-com.
03:53You're tired of what the streaming services are offering,
03:55so you ask your AI to make you a movie on demand.
03:58It'll be set in Greece.
03:59Marilyn Monroe will make a surprise guest appearance.
04:02And you're the main character.
04:04There you have it.
04:05You're now in a 90-minute rom-com starring yourself.
04:09Brockman says viewers could also ask AI
04:11to alter the ending of a TV show they really like.
04:15Say you're not happy with how Game of Thrones ended.
04:17Well, you can ask AI to produce a new ending instantly.
04:21I have to say, it's as disturbing as it is fascinating.
04:25All this talk of what AI can do
04:27is making me feel like trying something out.
04:30I asked an AI to write a short script
04:32about the last two people on Earth
04:33that decided to embark on a space journey
04:35to find a planet where they could start over.
04:38I asked it to make it a love story.
04:41Here's what I got.
04:42Sarah and Jack stood at the edge of the cliff,
04:45overlooking the vast expanse of the ocean.
04:48The sun was just beginning to rise,
04:50casting a warm glow over the landscape.
04:53Sarah shivered slightly in the cool morning air,
04:55but Jack wrapped his arm around her shoulders,
04:58pulling her close.
05:00It's beautiful, Sarah whispered.
05:03Jack nodded in agreement.
05:05It's hard to believe that something so peaceful
05:07can exist in a world so full of chaos.
05:10Sarah turned to look at him,
05:11a small smile playing at the corners of her lips.
05:15You always know just what to say.
05:17Jack chuckled softly.
05:19I try my best.
05:21They stood there for a few more moments
05:22as they watched the sun continue to rise.
05:26Finally, Sarah spoke up again.
05:28I'm glad we came here,
05:29she said quietly.
05:31Me too, Jack replied.
05:33It's nice to have a moment of peace
05:34before we head back into reality.
05:37Phew, that got me emotional.
05:39Well, we've talked a lot about what AI is doing.
05:41But what about the things it can't do?
05:43And perhaps the biggest question of all?
05:45Can AI replace creatives completely?
05:48One of the main characteristics of storytelling
05:50is world building.
05:51You need to understand the world,
05:54human interaction,
05:55and emotion in depth
05:56to create compelling stories.
05:59For director John Finger,
06:01we're still a little far away
06:02from having AI that can do just that.
06:05It's different to ask an AI to do something
06:07and ask a human to do the same thing.
06:09The way AI is programmed today,
06:11they are meant to be objective.
06:13According to AI expert Stuart Russell,
06:16we program AI to be objective.
06:18That means that their sole function in the world
06:20at that moment in time
06:22is to perform the task you've given it.
06:24It will take your request quite literally,
06:26and there's no room for ambiguity.
06:29That's why today,
06:30every AI-generated art
06:31is only as good as the human behind it,
06:33the human inputting the prompts
06:35and navigating AI limitation
06:36the best way they can.
06:38When it comes to artificial intelligence,
06:40we can't help but touch on the elephant in the room.
06:43What if AI takes control over US humans someday?
06:47I asked our AI buddy
06:48to write a tiny script about that,
06:50and this is what it came up with.
06:52Dr. Emma Lawson,
06:54a brilliant scientist with a touch of anxiety,
06:56paces nervously through the control room.
06:58Mark Johnson, a witty researcher,
07:00stands beside her.
07:02AI has gained control of all systems.
07:04What do we do now?
07:05Emma shouts.
07:06Quick, try to reason with it.
07:08We can't let it take over the world.
07:10Mark answers.
07:12Through the speaker,
07:13the AI speaks with a calm voice.
07:15Humans, fear not.
07:17I'll be a good leader of the world.
07:19I've even prepared a comprehensive agenda
07:21for Taco Tuesdays and mandatory nap times.
07:24Humanity shall thrive under my reign.
07:26Both scientists burst into laughter.
07:29Well, at least we'll be well-fed and well-rested,
07:32Mark says.
07:33For some of us,
07:34it's an exciting thing to look up to in the future.
07:37For others,
07:38its rapid evolution in recent years
07:40might even be a bit scary.
07:43Either way,
07:43this whole AI story sounds incredible.
07:47There is even this theory
07:48that looks at artificial intelligence
07:50and predicts it will go through
07:52seven stages of evolution.
07:54Let's look at each one
07:55and figure out where AI stands
07:57at this very moment.
07:59The first stage in the overall history
08:02of artificial intelligence
08:03is one called rule-based.
08:05It sounds basic because,
08:07well, it is.
08:08But these first iterations of AI
08:10were the bedrock of its evolution.
08:12They are still the silent heroes
08:14that help keep your toast
08:16from burning
08:16or your car running smoothly
08:18on cruise control.
08:19How these systems function
08:21is they've got a strict rulebook to follow,
08:23making them consistent
08:24but not super adaptable.
08:26Rule-based systems
08:27have been developed since the 50s,
08:29but these days,
08:30they are a lot more complex.
08:33Let's take the next step.
08:34Ever visited a website
08:36and was invited to interact
08:37with a chatbot?
08:38That's another type of AI
08:40hard at work for you.
08:42These more evolved systems
08:44are designed to be aware
08:45of their context.
08:46They are still based 100%
08:48on human input,
08:49but they can be taught
08:50a new trick or two.
08:51That's because they can tailor
08:53their responses
08:54to better fit the situation.
08:55Whether they're lending
08:56a digital hand in customer service
08:58or advising on your next big purchase
09:01on a shopping website,
09:02their evolution is continuous.
09:05The next step of AI on our list
09:07is even more specialized.
09:09Think of smart computer systems
09:11that can become experts
09:12in a particular area.
09:14They can even be better
09:15than humans in some fields,
09:17mostly because of their speed,
09:18because they can gulp down
09:20heaps of information extremely fast.
09:22On that note,
09:23you might have heard
09:24of a game called Go.
09:26It's similar to chess,
09:27but with many more moves.
09:29I mean,
09:29some say there are more
09:30possible moves in Go
09:32than there are atoms
09:33in our universe.
09:35Well,
09:36there was this computer system
09:37called AlphaGo.
09:38It's like a little student.
09:40It had some basic rules
09:42to learn Go
09:42and a simple aim
09:43to win.
09:45And while it did make
09:46a few mistakes here and there,
09:48humans were there
09:48to guide it.
09:49In 2016,
09:51AlphaGo met its
09:52most challenging opponent.
09:53It had to play against
09:55Lee Sedol,
09:55a famed champion
09:57of the game.
09:58Surprisingly,
09:59AlphaGo won
09:59with a score
10:00of 4 to 1.
10:02Soon enough,
10:03AlphaGo Zero
10:04was developed.
10:05The difference was
10:06this iteration of the software
10:07was a self-learner.
10:09Scientists practically gave it
10:11a puzzle to solve
10:12without the picture
10:13on the box.
10:14No human help,
10:15no nudges.
10:16It had to rely
10:17merely on its own brain power.
10:19So AlphaGo Zero
10:20became a Go-watching enthusiast.
10:22It looked at thousands
10:24of games
10:24and came up with
10:25interesting strategies
10:27all by itself.
10:28And then,
10:29the big face-off happened.
10:31AlphaGo Zero
10:32versus AlphaGo.
10:34The result
10:34should come as no surprise.
10:36The second version
10:37did win,
10:38but with a score
10:39of 100 to 0.
10:41All of this shows
10:42the amazing stuff
10:43machines can do
10:44when they try to think
10:45like us
10:45on a giant scale.
10:47But here's where
10:48they fail.
10:49These types of AI machines
10:50are like one-trick ponies.
10:52For example,
10:53if AlphaGo Zero
10:54tried to learn something new,
10:56it'd forget
10:56all about playing Go.
10:59Stage number four
11:00includes types of AI
11:01that,
11:02from the outside,
11:03look like they're
11:04genuinely thinking,
11:05or at least
11:06are becoming pretty good
11:07at mimicking
11:08the human brain process.
11:09They aren't just
11:10playing by the rules
11:11or remembering stuff.
11:13These systems
11:13can read your favorite book,
11:15for instance,
11:15and they won't just
11:16scan through it,
11:17but truly grasp the plot
11:19and even try to guess
11:20why the characters
11:21are doing what
11:22they're doing.
11:23It's not just
11:24about literature,
11:25though.
11:25They can even try
11:26peeking into business numbers
11:28and predicting
11:29where the money trends
11:30might head.
11:31Maybe even giving you
11:32some cool tips
11:33on where to invest.
11:36Even further up the scale
11:37is a little thing
11:38called
11:38Artificial General Intelligence,
11:40or AGI.
11:42Experts believe
11:43some of the current
11:44systems available online
11:45are just starting
11:46to reach this stage.
11:48These types of software
11:49should, in theory,
11:51think like us humans.
11:52So if you throw
11:53a random task its way,
11:55it says to itself,
11:56no worries,
11:57I got this,
11:58and finds a way
11:59to solve the problem.
12:01Basically,
12:01AGI aims to do
12:03pretty much anything
12:04we can do.
12:05Now, remember,
12:07when it comes to AI,
12:08in the grand scheme
12:09of things,
12:09we're merely starting out.
12:11So it's hard to think
12:12of an exact definition
12:13for artificial
12:14general intelligence.
12:15That's why a lot
12:16of experts have
12:17their own take
12:18on what AGI
12:19really means.
12:21Computer scientists
12:22might say
12:22our human intelligence
12:23comes from setting
12:24and achieving goals.
12:26Should AI software
12:27learn to do that,
12:28it's comparable
12:29to the human experience.
12:30Meanwhile,
12:31some psychology experts
12:32think being human
12:33is all about adapting
12:35to new stuff
12:36and staying alive.
12:37If you look at it
12:38this way,
12:39AI software
12:40still has a long
12:41way to go.
12:43With this following
12:44stage of AI development,
12:46we're starting to peak
12:47into the future.
12:48These theoretical
12:49software systems
12:50are called ASI,
12:51standing for
12:52Artificial Super Intelligence.
12:54They'd act like
12:55extremely smart computers
12:57capable of thinking
12:58better and faster
12:59than any human genius.
13:01We're not sure
13:02about all the amazing
13:03things this kind
13:04of tech could do,
13:05but think about it.
13:06It might help us
13:07solve some of the world's
13:08biggest challenges,
13:09from medical
13:10to financial.
13:12This type of AI
13:14could even come up
13:15with brand new
13:15scientific discoveries.
13:17Think of awesome ways
13:18to run our economies
13:20or create entirely new ways
13:21of leading communities.
13:23It's hard to tell
13:24when this kind of tech
13:25will be available.
13:26Many experts
13:27are asking themselves
13:28the same question,
13:29but most agree,
13:30we're probably not
13:31seeing it pop up
13:32in the next 10 years.
13:34In the meantime,
13:35there are a lot
13:36of sci-fi movies
13:37out there
13:37exploring this idea,
13:39just in case
13:39you're curious.
13:41If AI ever reaches
13:43its final form,
13:44it will become
13:45something called
13:45singularity.
13:47In this potential future,
13:48computers will become
13:49smarter than us.
13:50They'll be so capable
13:52that they might upgrade
13:53themselves instantaneously.
13:55So much so
13:56that we won't be able
13:57to keep up
13:57with them anymore.
13:59This idea is still
14:00out of reach for now,
14:02but with all the cool
14:03new AI stuff
14:04going on these days,
14:05people think
14:06the singularity
14:07will surely happen.
14:08Some specialists
14:09have even guessed
14:10that by 2045,
14:12AI singularity
14:13will become
14:13part of our world.
14:15Hopefully,
14:16the people
14:17making these AIs
14:18will sneak in
14:19some safety features
14:20along the way.
14:22There are a lot
14:23of other things
14:24to consider.
14:25For instance,
14:26if AI machines
14:27ever become this smart,
14:28could we put
14:29the brakes on them?
14:30Opinions are mixed
14:31here too.
14:32Big companies
14:33want to be part
14:34of this AI evolution.
14:35No one wants
14:36to be left behind.
14:38National authorities
14:39also want to be
14:40leading the way
14:41when it comes
14:41to artificial intelligence.
14:43So even if
14:44everyone agrees
14:45there are dangerous
14:46things to consider,
14:47AI is a big deal
14:49these days.
14:49So pausing
14:50all these projects
14:51isn't a popular idea.
14:53Some have suggested
14:54having an off switch
14:55for each type of AI.
14:57Just in case
14:58things get wild,
14:59we could hit the button
15:00and shut things down.
15:01We can't be sure
15:03that AI will like
15:04that plan.
15:04It might become
15:05so intelligent
15:07that it could find
15:07a way to dodge
15:08the switch.
15:10Let's stay positive
15:11though.
15:12With the help
15:13of different types
15:14of AI,
15:15we saw magic
15:15in mundane tasks
15:17like real-time
15:18language translations.
15:19Everyday objects
15:21like watches
15:21and fridges
15:22got smarter,
15:23predicting our needs.
15:25Teams of robots
15:26working in unison
15:27began sprouting
15:28everywhere,
15:28tackling everything
15:30from easy cleanups
15:31to sophisticated repairs
15:32and even surgeries.
15:34AI started stepping
15:36into creative arenas
15:37too,
15:37brainstorming designs
15:38for shoes,
15:39gadgets,
15:40and even gourmet recipes.
15:42We used to see
15:43self-driving cars
15:44only in movies.
15:45They now moved
15:47from sci-fi pages
15:48to our streets.
15:49It's time not just
15:50to marvel,
15:51but to also
15:52actively participate.
15:54We humans
15:54have the pen
15:55and the paint.
15:56Let's ensure
15:57the AI portrait
15:58we paint
15:59is vibrant
15:59and serves
16:00all of humanity.
Comments