Google Cloud Speech-to-Text API (Chirp) + Gemini 2.5 Flash Non-Thinking. Treat it as a searchable index of what was broadcast, not a quotation record.
00:00:00six World Cups, seven Copa Americas, world champion, twice chosen as the best
00:00:03player of this tournament, I'm talking about the World Cups and it is truly a
00:00:07cultural heritage of Argentina, so we will be here in CNN covering all
00:00:10of it. It will be until tomorrow. Hugo Mano Correa accompanied
00:00:22the world leader in
00:00:23news.
00:00:26Oppenheimer presents, brought to you courtesy of
00:00:41you just said, putting a
00:00:47Uade, the Argentine university that trains you to
00:00:50succeed in the world, with laboratories, technology and
00:00:53practices that propel you into the job market, study trips to
00:00:57the main cities of the world, you choose
00:01:01a pop-up appears where, when and how
00:01:02to study.
00:01:06Build your future at UAD, a great
00:01:09university. Why are people unhappy everywhere? Surveys
00:01:13show there is a wave of dissatisfaction and an increase in
00:01:17stress, sadness and frustration sweeping the
00:01:20world. In "How to Get Out of the
00:01:23are separate from Hole", journalist Andrés Oppenheimer shares innovative
00:01:26strategies to combat unhappiness. With interviews with
00:01:29the world's leading gurus of the new science of
00:01:32happiness, he tells us how countries, companies and
00:01:36people are using artificial intelligence and
00:01:39positive psychology to increase their life
00:01:43satisfaction. Now available in bookstores and online
00:01:46stores.
00:01:54Hello, I'm Endenheim. The topic of the moment worldwide,
00:01:58perhaps the number one topic is: the panic that has spread after warnings
00:02:02from several former employees of major tech companies that artificial
00:02:06day, that is, what intelligence could end humanity, not
00:02:10in 100 years, but in the coming years. Some have
00:02:13said that an artificial intelligence model could get out of
00:02:17control and hack an atomic power plant or
00:02:20biological weapons as early as in the next 10 years.
00:02:24So, the big question is, are these
00:02:27warnings realistic or are they exaggerated? And do
00:02:3212 a.m. and 6 those who make them have ulterior motives? It's not an academic question. The
00:02:37Independence at 9:30 in the panic broke out after a real event, when in
00:02:40July, artificial intelligence agents from OpenAI, the parent
00:02:43company of Chat GPT, penetrated
00:02:46the internet and hacked the tech company Hugging
00:02:49Face. OpenAI acknowledged the incident and said it would strengthen its
00:02:53security. But then, a few weeks ago, the alarms went off
00:02:57again when the former researcher resigned
00:03:04from his position citing fears that the artificial
00:03:07intelligence industry is in a race towards a super
00:03:10intelligence and that this will endanger our lives, he said
00:03:14accurate and timely information on on a social network that people
00:03:17control could mean the end of who are inside artificial intelligence companies believe
00:03:20that there is a great possibility that artificial
00:03:24intelligence, and I quote, "could
00:03:26kill us all" before the end of this...
00:03:30and I close the quote. And more recently, several
00:03:33executives and employees of these companies said that indeed there
00:03:37is fear within the industry that artificial
00:03:41you pay attention to and gives you a lot of intelligence will get out of control and cause a
00:03:44catastrophe. Today we will start the program by asking a
00:03:48well-known artificial intelligence expert if the
00:03:51end of the world is coming, if what they are saying is
00:03:55exaggerated, and then
00:03:58we will change the subject and talk with the Instagram and Facebook
00:04:02software engineer, who was a witness for the prosecution of several
00:04:05states in the recent lawsuit against Meta, the
00:04:09parent company of Facebook and Instagram. The
00:04:12lawsuit accused Meta of having intentionally designed its
00:04:16platforms to hook children and
00:04:20adolescents to screens, contributing to a mental
00:04:23health crisis among young people. The
00:04:26lawsuit, which was worldwide news, ended with an out-of-court
00:04:30settlement by which Meta will have to pay up to 18
00:04:33billion dollars and make changes to Instagram and
00:04:37Facebook to better protect minors. The company explicitly
00:04:40said it did not admit to having
00:04:43roulette, putting a gun to your done anything wrong, but as part of the agreement, it agreed to do several
00:04:46things, for example, set a 2-hour daily limit
00:04:50for screen time for minors under
00:04:5317 years old on its social
00:04:56comply with networks. But the big question is, will these measures be enough?
00:05:00Will the wave of anxiety and
00:05:03depression among youth be
00:05:06reduced with this? There are still many things to do. We will ask all this and much
00:05:10more to the former engineer. The former engineer who was the main
00:05:14witness against the company in the trial. Well,
00:05:18let's start with the first topic, the global panic about the possibility that artificial
00:05:21intelligence will end humanity in very few years.
00:05:25We have with us Stewart
00:05:29the Association for the Safe and Ethical Use of Artificial
00:05:31Intelligence and professor of computer science at the
00:05:35University of California at Berkeley. Let's go with him.
00:05:39Professor Russell, thank you for being with us. Well, let's get
00:05:42to the point. Do you think these warnings about
00:05:45a possible catastrophe due to artificial intelligence are
00:05:49justified or are they exaggerated? How
00:05:52concerned are you about all of this? First
00:05:56of all, I think it's important for people to know that these
00:05:59concerns are not new. The CEOs
00:06:02of major artificial intelligence companies
00:06:06have been saying this quite consistently,
00:06:09even before their own companies existed.
00:06:13So it's clear that it's not about inflating their stock
00:06:16prices, as some argue, because they were saying
00:06:19it even before they had stocks. The history of
00:06:22computing goes back at least to 1951.
00:06:26Alan Turing, founder of computing and one of the first to
00:06:30conceive of artificial intelligence, warned in
00:06:331951 that machines would likely
00:06:37take control if we managed to create machines more intelligent than
00:06:40humans, and the reason is quite obvious, right?
00:06:44Our intelligence gives us power over the world and all other
00:06:47species on Earth, and if we create systems more intelligent than
00:06:51us, then they will be more powerful than us, and how
00:06:55can we maintain power forever over entities much more
00:06:58powerful than us? Obviously, it's a difficult
00:07:02or impossible problem. This is basically what Alan Turing was telling
00:07:06us back in 1951.
00:07:08This is what the CEOs of the companies have been telling
00:07:12us. Dario Amodei from Anthropic, for example, estimated a
00:07:16risk of human extinction of up to 25%.
00:07:20So, there are two questions. Will we be able to create artificial
00:07:24intelligence systems with that level of capability, and do we have a
00:07:27solution to maintain control over them
00:07:30forever? The answer to the first question, most would
00:07:34say, seeing the pace of progress, is yes,
00:07:37we will be able to create systems with that level of
00:07:40capability. Do we want a solution to the control
00:07:44problem? The answer is no, companies openly
00:07:47admit that they don't know how to control even the systems
00:07:51they have now, much less systems that will be much
00:07:54more intelligent than human beings. You have
00:07:58been a skeptic about the proposals of Dario Amodei and others
00:08:02to slow down the race among companies
00:08:05towards artificial superintelligence. You said that the
00:08:08proposal to reduce the speed of this race
00:08:12is like going slower towards a
00:08:16cliff. Please explain that. Yes,
00:08:20well, given the argument I just made that creating
00:08:23machines much more intelligent than human beings and then losing
00:08:29the human race, if we are going full speed towards that
00:08:32goal, then we are going full speed towards a
00:08:35cliff, and saying, "Well, we're not going at 80
00:08:39km/h, but at 60 km/h," is very
00:08:42reassuring, right? The correct answer is:
00:08:46no, we have to go in another direction, and if we can't
00:08:49do that, we should stop. That's the same rule
00:08:52we use for airplanes, buildings, nuclear
00:08:56power plants, even for sandwiches. We say,
00:08:59there's a risk, either from a building collapse, a plane crash
00:09:03or food poisoning, and we demand
00:09:06certain safety tests before allowing a new type of airplane to be launched
00:09:10on the market and carry passengers.
00:09:42executives, Jacob, who resigned from
00:09:45Anthropic, and Evan Hubinger, the Anthropic executive who supported him,
00:09:48spoke of more than a 10% probability of human
00:09:51extinction. That's playing Russian
00:09:55head and pulling the trigger, hoping the bullets aren't in the right
00:09:59chamber. How can we allow this to be happening? Basically, let
00:10:03me see if I understood you correctly, you are saying that technology companies should first
00:10:08safety requirements and
00:10:11then, if approved, only then launch their
00:10:15products. You gave as an example the case of
00:10:17airplanes. Yes, I was questioning the idea that
00:10:21the solution will be to slow down the current race.
00:10:25Slowing down the process basically means, well, we're
00:10:28going to keep moving forward regardless of whether we can do it safely,
00:10:32we'll just do it at a slower pace, and I
00:10:35argued that if an aircraft manufacturer said, "We're going to introduce
00:10:39a new aircraft once a year and we hope that will give us enough
00:10:42time to carry out the flight tests and that the results will
00:10:46be good, but we're going to introduce the aircraft anyway," if that were the
00:10:50case, the government would say,
00:10:53"No, you're getting it completely wrong, you can't
00:10:56launch aircraft whenever you want, you have to do the flight tests
00:11:00and the results of the flight tests have to be good, and the same goes for the
00:11:04structural integrity and the tests of the wings and engines." There
00:11:08are regulations that say they have to comply with all these safety
00:11:12standards, because otherwise they are putting people
00:11:15at risk and we will not accept it.
00:11:18Many ethics experts advocate for
00:11:20labeling. They say that artificial intelligence companies and social
00:11:25networks should be required to label their
00:11:28products, just like supermarkets do with
00:11:31food. That is, for example, to indicate if a video or an image, a
00:11:35photo has been created with artificial intelligence, or what
00:11:38percentage of the content of a video has been
00:11:42created with artificial intelligence. What do you think about
00:11:45that? How would it work? This is already a requirement in Europe and
00:11:48China, where there are very strict regulations on the labeling of
00:11:52artificial intelligence content. I think that
00:11:55it actually works. It is curious that American
00:11:58companies argue in the United States that
00:12:02this is too burdensome, "we cannot comply with these terrible
00:12:06regulations proposed by the states of the United States," while
00:12:10they comply with them in Europe and have succeeded without problems, and it is not that
00:12:13difficult. I think the simplest example is that one should
00:12:17have the right to know if they are interacting with a human or with a
00:12:21machine. It is very simple, if it is a dialogue, either.
00:12:30That would include, for example, when one receives a tweet or a message on social
00:12:34networks, on Twitter, on Facebook or on any other social
00:12:37network. Should social networks explicitly
00:12:41notify us that it is a bot? Of course
00:12:45they should, and I should have the option to filter responses or
00:12:49friend requests from bots. Personally, I don't
00:12:52want to be friends with a bot, I don't want bots in my
00:12:55inbox with nonsense generated by other bots. I should be able to
00:12:59filter non-human interactions.
00:13:02Professor, all this debate about the potential
00:13:05apocalypse due to artificial intelligence is overshadowing another
00:13:09debate, the debate about the effect of artificial intelligence on our
00:13:13minds, right? What experts call cognitive
00:13:16surrender. What advice would you give to our viewers
00:13:20on how to protect their minds, their brains from this
00:13:23bombardment of messages and artificial intelligence algorithms
00:13:26that manipulate our thoughts? How can we preserve
00:13:29our mental independence?
00:13:33It's a very important question, and there is
00:13:37increasingly clear evidence that using artificial intelligence to
00:13:40perform cognitive tasks results, as you say, in
00:13:44cognitive surrender, cognitive deficit, cognitive
00:13:47decline. We really need to find an appropriate
00:13:51term for this. Basically, one becomes
00:13:54dumber. One becomes less intelligent, loses their
00:13:57mental capacity. For example, there are people who cannot
00:14:01answer a question without first consulting their artificial
00:14:04intelligence platform. They cannot organize their daily lives, they cannot
00:14:07work. I have spoken with managers whose direct
00:14:10subordinates cannot function without constant help from the artificial
00:14:14intelligence system. We have very clear evidence of an
00:14:17experiment conducted in China with 26 children. Some
00:14:21used artificial intelligence to help them with their homework and others
00:14:25did not. The children who used artificial intelligence scored
00:14:28higher on their homework and lower on their
00:14:32final exams, closed-book final exams
00:14:35and without artificial intelligence. Therefore,
00:14:38the use of artificial intelligence clearly resulted in a
00:14:41failure to learn. So, I think that as
00:14:45this becomes more widespread, it will be increasingly inexcusable that
00:14:49we allow children to continue using artificial intelligence to
00:14:52copy and paste their homework and substitute
00:14:56think. It would be a total negligence on our part, and it
00:15:00is already happening. Professor Russell, thank you very much, we have
00:15:03a break now. When we return, we will talk to the former
00:15:06Instagram software engineer,
00:15:09who was a witness in the trial against the company for allegedly damaging
00:15:13the mental health of minors. Don't go away, we'll be
00:15:17right back.
00:15:18volvemos,
00:15:24There are places that... stay with you forever, and that is Dominican
00:15:28Republic. Dominican Republic is beach and
00:15:31mountain, it's city, countryside and a path to discover. It's the
00:15:35flavor of a cuisine born from the land. Music that is danced
00:15:38and felt, it's people who receive you as if they had known you for
00:15:42years, hands that create, communities that share,
00:15:45stories that make you part, it's a country that grows
00:15:48caring for what makes it unique, Dominican Republic
00:15:51feels. Feel Dominican Republic. The full story
00:15:55with Anderson Cooper, CNN's best journalists share deeply
00:15:58documented stories.
00:16:00documentadas.
00:16:07your advice be to me, you believe the civil
00:16:10government should be guided and dictated by biblical law, how
00:16:14in the world did you survive
00:16:15this?
00:16:19The full story with Anderson Cooper, Sunday on
00:16:23CNN. Do you have 5 minutes? We have five
00:16:27things for you, a podcast with the five most important topics
00:16:31of the day. Just ask your smart device, "Give me
00:16:34the latest news from CNN." Next, I'm
00:16:37going to tell you the five things you need to know at this time, updated every
00:16:41hour, free of charge or commitment. Stay informed
00:16:45from anywhere and at any time. Subscribe
00:16:48to our newsletter. Listen to the five things you need to know
00:16:52throughout the day.
00:17:35The need to migrate knows no age. Many
00:17:39say they want to reach Mexico City, but others are looking
00:17:42to go further
00:17:43north.
00:17:47The Mexican army's anti-drone battalion consists of
00:17:50these types of elements, jammers to bring down drones
00:17:54that have become a threat. The
00:17:57Chiapas authorities have detected a cattle smuggling
00:18:00point. This is one of the illegal crossing
00:18:04points for cattle going from Central America to
00:18:07Mexico. Thousands of demonstrators, workers from this
00:18:11country, left the Angel of
00:18:14morning. I am in the municipality of Teuchitlán, in the
00:18:18Mexican state of Jalisco. Here the discovery was
00:18:21macabre by collectives of people searching for the
00:18:25disappeared. What made you come here with the suspicion
00:18:29that your son might be on this
00:18:30ranch?
00:18:44Thank you for staying with us. A few weeks ago, Meta, the parent
00:18:48company of Instagram and Facebook, reached an out-of-court
00:18:51settlement with 48 states of the United
00:18:54States, for which the company will have to pay up to
00:18:5718 billion dollars and make several
00:19:00changes to its platforms to
00:19:03better protect minors. The lawsuit had accused
00:19:06Meta of intentionally causing screen addiction
00:19:10in children and adolescents,
00:19:13contributing to a mental health crisis among
00:19:16young people. The company has explicitly denied
00:19:20having done anything wrong, and as part of that agreement, it agreed to
00:19:24several demands, such as, for example, setting a 2-hour daily
00:19:28limit for the time minors under
00:19:3117 years old can spend on its
00:19:34platforms. But will that be enough
00:19:37to combat this wave of anxiety and youth depression that
00:19:41we were talking about? We have with us the man who was a witness
00:19:44against the company in the trial, former Instagram software
00:19:48engineer Arturo Bejar. Let's point out
00:19:51that we invited Meta three weeks ago to interview an official
00:19:55spokesperson for the company, but so far we have not received a
00:19:58response. Instead, they sent us the following
00:20:01statement, uh, I quote, "Meta has reached an agreement
00:20:05with a bipartisan group of... attorneys
00:20:08general, uh, state attorneys general across the country,
00:20:12on a new set of rules governing adolescents' use of social networks.
00:20:16The framework we have negotiated will allow
00:20:19parents to easily manage their children's access
00:20:22to our platforms. Our new commitments on time limits, night mode features,
00:20:26and usage limits during school hours set the right path for the entire
00:20:30industry. But this framework will only work if all our
00:20:34competitors join in. Given that adolescents frequently use
00:20:37dozens of applications, we need an industry-wide
00:20:41solution." I close the quote. Let's go with Arturo
00:20:45Bejar. Arturo, thank you for being with us. Well. As we just
00:20:49said, you were a key protagonist in this trial. What is
00:20:52your conclusion after this agreement? Will these measures be
00:20:56enough to protect minors?
00:21:00No. It's a very important change that the
00:21:02states achieved in the agreement, because there are
00:21:06some good things, but at the same time it's a start and it's
00:21:10not enough. It's like telling our children that they
00:21:14can smoke all the cigarettes they can smoke in 2
00:21:17hours. To change the nature of cigarettes to
00:21:20make young people safer on these products, let's
00:21:24break it down. One of the measures that Instagram and Facebook
00:21:28committed to was what
00:21:312-hour daily limit on screen
00:21:35time, on the time minors can spend on their
00:21:38platforms to precisely reduce
00:21:42screen addiction. How will that work? That
00:21:45means that after 2 hours the transmission stops or
00:21:51saying: "We advise you not to continue." How will this
00:21:54work? How it should work is that after 2
00:21:57hours, when one tries to open Instagram or anything
00:22:01like that service on our phone, it should be a
00:22:04screen saying: "Today the 2 hours are up and
00:22:08nothing else can be done except messages," because messages
00:22:13what is the normal use of the social network. Two hours per
00:22:17day or 2 hours in a row? It's 2 hours per
00:22:20day, and in theory, at 15 minutes, they will give a
00:22:24warning, but these would have to be made in such a way that they were
00:22:27effective for something that, when you see
00:22:31it, makes you put down the phone, not just a small
00:22:35screen in between that everyone says yes to and then continues
00:22:38doing what they were doing. Do you think this 2-hour
00:22:42daily limit will make a difference, that it will be important?
00:22:46It can make a difference,
00:22:49but the... the problem still is that for the other years that young people
00:22:53live on these products every
00:22:56these social networks are giving our children, that is not
00:23:00changing. So it's good that there is a
00:23:03limit of fewer hours a day, provided it is
00:23:06done in a way that is effective, but something
00:23:09needs to be done about the other harms. Another measure agreed upon by Instagram,
00:23:13Facebook and the plaintiffs was that minors will no
00:23:16longer be able to use Instagram and Facebook between
00:23:22a.m. to prevent sleep deprivation, which
00:23:26according to many psychologists causes anxiety and depression. Will those
00:23:296 hours of blocking be enough to
00:23:33reduce mental problems in minors? We hope it
00:23:36helps, and the fact that they are doing it is very
00:23:39important, but something that is missing in this
00:23:43issue, which is necessary for governments in the United
00:23:47States and Latin America and other places to do,
00:23:50is to ask young people how much these changes are
00:23:54helping them. But if you ask young people, "Hey, how's
00:23:58your sleep, what product are you using, what's keeping you
00:24:01awake?", they will tell you, and at the end of the day,
00:24:04results are more important than efforts. What did
00:24:07Instagram and Facebook agree with the plaintiffs on the
00:24:10issue of infinite scrolling? That is,
00:24:14infinite scrolling. With
00:24:17respect to the addiction mechanisms, such as
00:24:21infinite scrolling, videos that start playing
00:24:24automatically, notifications,
00:24:27an algorithm that notices what
00:24:31that, whether it's good or bad for you, in those areas there were
00:24:35no agreements. What was changed is how much time. So
00:24:39that is a job that needs to continue, and what we need to
00:24:42know, both parents and governments, is that this is the beginning of a
00:24:46process, not the end. Let me see if I understand correctly what you are saying:
00:24:50the usage time was changed, but not the content.
00:24:54Exactly, the time is limited, but the things
00:24:58that we are learning and know cause harm, which is infinite
00:25:01scrolling, I mean, from the moment you are born until you die, you can be
00:25:04watching, the videos that start to
00:25:08play before you say, "I want to
00:25:12see that," and companies know and
00:25:15studies show that when you you see a video,
00:25:19you immediately prepare to pay attention to it and see what is
00:25:22happening, it's natural, you enter a bar or whatever, there's a
00:25:25screen, our eyes immediately spend a few
00:25:29seconds there to see what happened and when that video has
00:25:32something so strong, something exhausting,
00:25:35something that affects you, well, you've already experienced it, right?
00:25:39when you saw it, let's remember that several weeks ago we invited
00:25:42Meta, the parent company of Instagram and Facebook, to participate in today's
00:25:46program, but we did not receive a response. Previously, Meta had
00:25:50denied that its platforms harm minors and said that the
00:25:53agreement reached with the plaintiffs does not imply that
00:25:57it admits to having done anything wrong. The company sent us a statement from its chief
00:26:00legal officer, CH Mahone, saying,
00:26:03and I quote, "Meta has reached an agreement with a bipartisan group of
00:26:07state attorneys general across the country on a new set of rules
00:26:10regulating the use of social networks by
00:26:13teenagers. The framework we have negotiated will allow parents to easily
00:26:17manage their children's access to these platforms.
00:26:21Our new commitments on time limits, night mode features,
00:26:24and school-day usage limits mark the right
00:26:28path for our entire industry. But this framework will only
00:26:31work if all our competitors join. Given that
00:26:34teenagers frequently use dozens of applications, we need
00:26:38an industry-wide solution." End quote.
00:26:41We have to not cut when we come back, we'll continue with the
00:26:44witness in the trial against Meta. Don't go anywhere, we'll
00:26:48be back.
00:26:53Every night there is the possibility of an
00:26:57unfiltered bombardment. Our young people should
00:27:00not be treated like this anywhere. Here we
00:27:04ask, is there freedom of belief in the country, in Spain? One
00:27:08hears that this type of phenomenon is not that they are the
00:27:11generators of change, they are the expression of change, and it is debated, will
00:27:15your foreign policy be like this, lukewarm, without taking
00:27:18positions? Conclusions with Fernando del Rincón from Monday to
00:27:21Friday on
00:27:22CNN.
00:29:26City,
00:29:28a story can change everything, a
00:29:31look. at the reconstruction of a city and the
00:29:34hope of a people determined to rise again, it evolved very
00:29:37quickly and we never thought it would become a category
00:29:41five. And to understand what is happening, we go
00:29:44to my colleague Ione Molinares in Washington,
00:29:48the explosion, we can see the dimension it had to take down the entire
00:29:52facade. We hear the protagonists, they
00:29:56took care of us as mothers and fathers of the
00:29:5943, where she lived, a bomb fell. We are about
00:30:02to board this Mexican Air Force plane bound for
00:30:06Romania, to have a complete view. We are
00:30:09here as immigrants to progress, to make this country grow. There are
00:30:13great stories to know and many perspectives to analyze.
00:30:17Perspectives with Mario González, Monday to Friday
00:30:20on CNN.
00:30:53Hello, I'm Andreas Openheimer. If you liked today's program and want to recommend it to your
00:30:57friends, you can find it on YouTube.