- 21 minutes ago
The China Now special program informs about this country's news. Telesur English presents a new episode of “China Now”, a wave media's production that showcases the culture, technology, and politics of the Asian Giant. China Now is a show that explores the past and future of the Asian Giant
teleSUR
These and more stories now! 🌐Join teleSUR's community and be part of a space where your ideas matter. https://chat.whatsapp.com/KLABTE41aSj3ioYabXQQ6K 🗣 Don't be left behind, join teleSUR's Channel https://whatsapp.com/channel/0029Va98Z437Noa2tjytCt1N Web💬 https://www.telesurenglish.net/ X🧶 https://twitter.com/telesurenglish Instagram 📸 https://www.ins tagram.com/telesurenglish/?hl=es Facebook 📱 https://www.facebook.com/teleSUREnglishOfficial Youtube ⏯ https://www.youtube.com/channel/UCmuTmpLY35O3csvhyA6vrkg Tiktok👌🏼 https://www.tiktok.com/@telesurenglish?_t=8hMLCY0a6z7&_r=1
teleSUR
These and more stories now! 🌐Join teleSUR's community and be part of a space where your ideas matter. https://chat.whatsapp.com/KLABTE41aSj3ioYabXQQ6K 🗣 Don't be left behind, join teleSUR's Channel https://whatsapp.com/channel/0029Va98Z437Noa2tjytCt1N Web💬 https://www.telesurenglish.net/ X🧶 https://twitter.com/telesurenglish Instagram 📸 https://www.ins tagram.com/telesurenglish/?hl=es Facebook 📱 https://www.facebook.com/teleSUREnglishOfficial Youtube ⏯ https://www.youtube.com/channel/UCmuTmpLY35O3csvhyA6vrkg Tiktok👌🏼 https://www.tiktok.com/@telesurenglish?_t=8hMLCY0a6z7&_r=1
Category
🗞
NewsTranscript
00:08Hello, Telesur English presents a new episode of China Now, a Wave Media's production that
00:13showcases the culture, technology and politics of the Asian giant.
00:17In this first segment, China Courants take a deep dive into the week's top stories from
00:22Anthropik accusing several Chinese companies of trying to extract capabilities of clod
00:27to the United States, continuing to emphasize the importance of keeping the America ahead
00:32of China in AI.
00:33Let's see.
00:49Hello, and welcome back to China Currents.
00:51On September 10th, Anthropik accused several Chinese AI companies of trying to extract capabilities
00:56from clod, and the numbers are huge.
00:59Anthropik says Alibaba alone made more than 151 million exchanges with clod between May
01:04and July using thousands of accounts.
01:06It also accused DeepSeek, Moonshot, Jipu and Xiaomi of doing something similar, collecting
01:12clod's outputs to improve their own models.
01:15Washington is now calling this industrial-scale AI theft.
01:19But Anthropik's report raises a bigger question.
01:21Who gets to define AI theft?
01:23Because the controversy here is not exactly about the distillation itself, because distillation
01:28is a standard technique in AI development where one model learns from the outputs of another.
01:34And Anthropik itself acknowledges that frontier AI companies routinely use distillation to create
01:39smaller and more efficient models.
01:41The controversy is about where that normal technique crosses into unauthorized views.
01:46Anthropik says some Chinese companies used fake accounts, proxy networks, and other methods
01:50to get around its restrictions.
01:52Anthropik calls these operations illicit distillation.
01:55But where exactly is the line between learning from a competitor and stealing from one?
02:00What was clod itself trained on?
02:02And where does the knowledge inside these AI models actually come from?
02:06AI companies don't create all that knowledge from scratch.
02:09They learn from information created and accumulated by the world.
02:13So when one AI learns from another, is that really stealing knowledge?
02:17Or is it learning from a technology that was itself built on humanity's existing knowledge?
02:22And there's another difference that often gets overlooked.
02:25Many Chinese AI models have embraced open source, allowing others to study, modify, and build on them.
02:31Some of the most powerful US models, by contrast, remain behind closed doors.
02:35If AI is going to reshape the world, should its most powerful capabilities belong to a handful of closed companies?
02:41Or should they become increasingly accessible to the rest of the world?
02:45China's commerce ministry has rejected the allegations, calling distillation a widely used technology
02:50and accusing Washington of turning a technical and commercial issue into a tool for restricting competition.
02:57At the same time, the US government is increasingly treating AI competition with China as a national security issue,
03:02as Chinese companies develop increasingly competitive AI systems,
03:06the ways they can gain access to frontier technology are becoming part of that same security debate.
03:11And that changes the context of the argument.
03:14If access to the frontier is restricted, competition does not simply stop.
03:18Companies look for other ways to learn, develop, and close that gap.
03:22That is where distillation becomes particularly important.
03:25A model can be closed, its weights may be protected, its training data may be secret,
03:29but its capabilities are visible every time someone interacts with it.
03:34And those interactions can become a source of training data.
03:37There are also signs that the story of AI competition may be more complicated
03:40than a simple China-copied-America narrative.
03:43DeepSeq's R1 research, later published in Nature,
03:46describes a training process centered on reinforcement learning and its own model development pipeline.
03:51And Kimi K3 appeared only about two weeks after Anthropic released its Fable model.
03:57Some AI researchers questioned whether there was even enough time to collect sufficient data,
04:02train a model, and reproduce that level of capability through distillation alone.
04:06These cases do not settle the broader dispute,
04:08but they do highlight a basic problem with measuring AI progress.
04:12It is becoming increasingly difficult to identify exactly where a model's capabilities come from.
04:17A model can learn from many sources,
04:19and researchers can combine different datasets, training methods, and models.
04:24Capabilities can spread through the industry much faster than companies developing them can control.
04:28In the past, controlling a technology often meant controlling the hardware,
04:33the patents, or the manufacturing process.
04:35But with AI, the line is much blurrier.
04:37You can restrict the chips, and you can restrict access to the model,
04:41you can restrict the API.
04:42But once a powerful model is out in the world,
04:45controlling everything others can learn from becomes much harder.
04:49And that might be the real challenge for Washington.
04:52Not simply whether China is learning from American AI,
04:55but how quickly that learning can turn into competition.
04:57The story becomes more complicated.
04:59Just days after releasing its threat report, Anthropic CEO Dario Amadei called for slowing the
05:05development of increasingly powerful AI, warning about potentially catastrophic risks.
05:10Other major AI leaders, including OpenAI CEO Sam Altman and Elon Musk, have also called for greater caution.
05:17At the same time, Washington continues to emphasize the importance of keeping America ahead of China in AI.
05:23So, two priorities are now running in parallel.
05:26One is to control the risks of increasingly powerful AI.
05:29The other is to make sure America remains ahead in the global AI race.
05:33Which brings us back to the question at the center of the story.
05:36Who gets to decide the rules of AI competition?
05:39If AI's progress is built on the knowledge and work of previous generations,
05:44then the bigger question may not be simply who copied whom.
05:47It is whether the progress of AI should be controlled by a few companies,
05:52or shared by humanity as a whole.
05:54Now, let's turn to our second story.
05:56One about diplomacy.
05:57China's President Xi Jinping is back in India for the first time in seven years.
06:01But this trip is more about than one summit.
06:04The Chinese president joined leaders from 11 BRICS countries in New Delhi with wars,
06:10trade tensions, and the future of global technology all on the agenda.
06:14Before the formal summit, Xi met with Indian Prime Minister Narendra Modi for talks covering trade,
06:19the border, supply chains, and people-to-people exchanges.
06:22Both sides agreed to keep expanding trade and business ties, while continuing talks on their border dispute.
06:27They also discussed restoring more direct links between the two countries, including travel and other exchanges.
06:34The China-India trade hit a record $155.6 billion last year.
06:40So while the two countries compete in many areas, their economies are becoming increasingly difficult to separate.
06:47The focus shifted then to the wider BRICS group.
06:50The bloc has now 11 members, including China, India, Russia, Iran, Saudi Arabia, the UAE, Indonesia, Egypt, Ethiopia, Brazil, and
07:00South Africa.
07:02This year's summit comes as several of its members are directly affected by the war in the Middle East.
07:07BRICS leaders call for maximum restraint and a political and diplomatic solution to the conflict.
07:12But the summit wasn't only about war and security. Technology was another major focus.
07:18China proposed deeper cooperation among BRICS countries on artificial intelligence,
07:23including a BRICS AI open-source community and a digital ecosystem platform.
07:27The idea is to expand cooperation and access to AI technology across the group.
07:33Finance was another big item. BRICS members agreed to keep working on cross-border payment systems
07:38and to expand the use of local currencies in trade and investment.
07:42But despite years of speculation, there is no new BRICS currency announced.
07:47Instead, the focus is on making different payment systems work more easily across borders.
07:51And there was one more important piece of news announced at the summit.
07:55Next year, the BRICS presidency moves to China.
07:58So this year's meeting in India was not just another gathering of emerging economies.
08:03The group is now bigger, its agenda is broader, and its members are trying to coordinate on everything
08:09from AI and payments to trade and global security. And next year, China will be the one setting the agenda.
08:16Now, let's turn back to the domestic front. A fire abroad, a foreign flag cargo ship in China,
08:21has killed 25 people. The fire broke out on September 10 aboard the Ocean Melody,
08:27a 20-year-old bulk carrier undergoing maintenance at the Qingdao Beihai shipyard.
08:31The ship sails under a Liberian flag. It arrived at the shipyard on August 31 for repairs and inspection.
08:38It is managed by a Shanghai-based company, while the shipyard is operated by China's
08:43State Shipbuilding Corporation, one of the world's largest shipbuilders.
08:47There were 42 people on board when the fire started at around 11.15 in the morning.
08:5112 people were safely evacuated, while another 5 were injured and taken to hospital.
08:56The remaining 25 were later found dead.
08:59So, what happened inside the ship? The fire broke out while the vessel was
09:03undergoing repairs and inspection, and it took firefighters more than three hours to fully
09:07extinguish the blaze. When investigators entered the damaged areas, they found a welding torch on
09:12the second level below deck, along with large amounts of oil residue on the floor. The fire damage
09:18was also much more severe on the lower levels, with firefighters reporting that the flames appeared to
09:23spread upward from the bottom of the ship. They also reported a strong smell of diesel inside the vessel.
09:29These findings have raised the possibility that hot work, such as welding, may have ignited flammable
09:34oil or vapors inside the engine room. But that is still only a preliminary line of investigation.
09:40The official cause of the fire has not yet been announced. And because 25 people died,
09:44the investigation is now going far beyond simply finding where the fire started.
09:49China's State Council has launched a national-level investigation involving multiple
09:53government departments. For now, investigators will need to determine not only what started the
09:58fire, but also whether there were failures in safety management and how a ship undergoing routine
10:03repairs ended up becoming a deadly trap for so many people. But not all of this week's ocean news was
10:08bad news.
10:09Chinese scientists say they have discovered a major mineral deposit deep beneath the Pacific Ocean.
10:14Samples from the site contained up to 15.4 grams of gold per ton of ore, while silver concentrations
10:20reached as high as 1,271 grams per ton. The deposit lies in the western Pacific, about 1,200 to
10:291,500
10:30meters below the surface within China's exclusive economic zone. The discovery came during an 18-day
10:36expedition aboard the research vessel Xiangyang Hong-10, with scientists from Tsinghua University
10:43and several other institutions conducting seven deep-sea dives. But this isn't simply an underwater
10:48gold mine. The site is an active hydrothermal field where mineral-rich fluids rise from deep beneath
10:54the seafloor and deposit copper, zinc, gold, and silver as they mix with cold seawater. Researchers say
11:00the gold and silver concentrations are significantly higher than those typically found in land-based deposits.
11:06However, there is a major catch. The gold and silver found are the highest concentrations found among
11:13the samples, but not the average across the entire deposit. And despite the impressive numbers,
11:19the site cannot simply be mined. At more than a kilometer below the ocean surface, extreme pressure
11:25and complex geological conditions makes deep-sea mining extremely difficult. For now, the discovery
11:32is less about mining gold and more about mapping what lies beneath the Pacific. China has already
11:38conducted more than 10 surveys in the region and identified dozens of high-temperature hydrothermal vents.
11:43This latest expedition adds another potentially valuable deposit to that map, and it shows just how
11:49much of the deep ocean remains unexplored. And that's our four stories for the week. Thank you for
11:55watching China Currents, and we'll see you next time.
12:09We have a short break now, but don't go away because we'll be right back.
12:24Welcome back to China Now. In this second segment, Overlap has Jeff Xiong, the Secretary General
12:29of the Global South Academic Forum, and also an AI expert going deeper on this topic. Let's have a look.
12:54Hello and welcome to one more episode of Overlap, this conversations from across the world. Tonight,
13:01we have the great pleasure of being joined by Jeff Xiong. He is the Secretary General of the Global South
13:08Academic Forum and also an AI expert. And we're here to have an important conversation, such a pressing
13:15conversation about AI at this moment. Thank you, Jeff, for joining us here in Overlap.
13:20You're welcome. Happy to have this opportunity.
13:23Of course. And you know, we use this time in order to get a little bit deeper into some of
13:28the most
13:29pressing issues and bringing this perspective from different sides of the world. And of course,
13:35China is such an important partner in this moment in order to understand AI development. And
13:41I just wanted to start this conversation by thinking about how polarized the discussion about AI seems to
13:49be right now. For some, it seems that AI is going to solve all productivity issues and will bring
13:57all solutions we haven't found yet. And for others, it's the great enemy, the great danger that is
14:03putting really humanity at risk. On that spectrum, where should we put AI right now, particularly from the
14:13Global South? How would you start that conversation? I think I can understand the concerns, the worries
14:20about AI. Of course, first of all, we all know the potential. It can accelerate the speed of science and
14:30social science research. It can help people do many things. Robots can perform difficult physical work
14:40for people. But I totally understand people also have the worries, especially if you look at what AI,
14:49particularly in the United States, what AI being used in last year. It's been used for wars. It's been
14:59used to monitor and surveil people. And companies are giving their information about individuals and
15:08organizations to intelligence organizations like CIA. And they don't even hide that. And the building,
15:18the expansion of data centers are leading to environmental and ecosystem worries. And of course,
15:28there is also a worry to employment. So I think it's people have good reason
15:36to be worried about AI. But I would say that is more a political issue than a technical issue.
15:44It's about how the technology is being used and how it's being regulated. And that is a political
15:53decision. Well, you can say in some cases, in some countries, there is a lack of political guidance.
15:58There are companies making their decisions without political regulation. And that, per se, is also a
16:07political decision. So I think what we're seeing, the increasing worry of AI development showing us
16:18the lack of a good political guidance in some parts of the world, especially in the United States.
16:26So then the question to the global south becomes, is that the only approach, the only solution we
16:35could ever have? Based on that point of view, we would say that's why the experience of China, the model
16:45of China becomes so important because it shows there is not only a single approach for AI development. There are
16:55alternatives. At least there is another alternative, which is a socialist
17:01approach developing AI. Exactly. So you were touching on some very important topics. And first of all, because we're trying
17:08to
17:09bring this argument in order to make for a better well thought discussion of these topics online for a general
17:19audience and a general public who are a lot very concerned about everything that has to do with the AI
17:25right now.
17:26When you're talking about the open source model, how would you explain that for someone who has
17:35no idea in terms of what goes into in terms of the the production and the programming side of the
17:45AI?
17:46How would you describe what an open source model is that you are linking to a Chinese way of developing
17:53AI?
17:54And what is it pulled against? What would be the alternative?
17:58Interesting. That's it. Thank you for giving me the interesting challenge.
18:04So in many ways, AI we're talking about today are very similar to to human's brain, right? Human's brain have
18:15our brain is basically a neural network network with hundreds of billions of neurons connecting to each
18:24other. And that connection make what intelligence possible, right? So AI we're talking today,
18:33in many ways are very similar to that structure. So there are hundreds of millions, billions,
18:40sometimes already trillions of neurons connecting to each other. They are not biological neurons,
18:48they are parameters. So they are parameters in computer, in data, in computer science terms.
18:55And you can imagine when a company first developed such an advanced AI, such an advanced model,
19:06it can choose to close that model, to hide that information from anyone else.
19:15They hide the parameters so that they can monopolize the intelligence. Then if they are the only company
19:25can do that, they can basically decide how much premium they want to, how much profit they want to earn
19:31from the monopolization, right? This is economic 101. And that is exactly what OpenAI and Google
19:40and Anthropoc are doing. You don't see they have open source models. They hide the parameters. Only they
19:47themselves know how their brains are made. The Chinese companies, led by DeepSeq and followed by other
19:58companies, selected a different approach. They decide when they developed an advanced AI, an advanced
20:05artificial brain. They open all the parameters. So it's a big file. It's going to be a few gigabytes to
20:14a dozen, dozens of gigabytes size. But in theory, anyone in the world can just download the parameters
20:25and replicate the model on their own machines. It's not cheap machines, okay? It's still expensive machines,
20:33but they don't monopolize it anymore. Other companies like OpenRouter, for example, or Alibaba. They can
20:41just download those models and host them. And so that they start to offer the same level of capacity
20:49to the world. Of course, you can imagine a country, for example, Venezuela government can decide,
20:57let's build a data center and host that model in the data center so we can offer, even though the
21:04country is being sanctioned, but we can still offer the capacity to our people. Or a university can do that
21:13as well. So they don't have to rely on any particular company to offer them the computation power to use
21:24the intelligence. And more importantly, they are not bonded anymore to the monopolization.
21:31The company cannot get extra high
21:35profit out of them because they have choices now. So that is when I say
21:42an open ecosystem, that is what it means.
21:45In such an open ecosystem, you have a few different models at least. As of today, you have DeepSeq,
21:54you have GLM, you have Kimi, and QN. They're all pretty good. I'm not saying they are the best today.
22:02They are a few months after the best, but they are pretty good. And with that options,
22:11an organization can choose from those models and decide, okay, this one fits me better. And six
22:18months from now, if this model lagging behind, we can choose to another one. They are not monopolized
22:26by any particular vendor anymore. So in my concept, I call it the commoditization
22:38of larger language models. Basically, you make AI a commodity. AI is high technology, is advanced
22:46technology, but it doesn't have to be super expensive. It can be a commodity just like
22:51electricity. Electricity is also high technology, but it's not super expensive. The same to artificial
22:58intelligence. You were talking about the possibility of, for example, Global South Latin American government
23:05is open right now. Consumption of having the decision of taking this open source models and using them to
23:13to foster their own capacity and their own challenges. What do you see in terms of the great opportunities
23:24that AI taking in this way could open, particularly for Global South countries like Latin American and Caribbean
23:33countries open right now? Considering the great dependency that we have of general economic structures,
23:41the gap differences, and the technological gap that is always putting this sector, this Global South, just behind
23:53general global supply chains and everything. Is there a key there in AI that we could tap into?
24:01You know, in the last 20, 30 years, there has been always a structural insufficiency
24:13of supply of software to the Global South. In many cases, Global South users are not only relying on the
24:25Global North for infrastructures, but also relying on the Global North for ideas of what software should
24:32be built and what users should be served. We have to, in many cases, we have to follow the Global
24:39North's
24:40model to decide, okay, because there are existing software and there are software companies building those
24:48things. So we better adjust our operation model and management model so that we can benefit from using the software.
24:59It happened in the 1990s when ERP getting popularized. It happened in the 2000s, Web 2.0 being popularized, and
25:09then when people talk about Web 3.0 and
25:14the decentralized internet, it's always the Global North had some ideas. Let's be more clear, it's almost always
25:28Wall Street and the Silicon Valley had some ideas of how information should be used and managed
25:36in the society. And then Global South follows those ideas and those software. Because we didn't have that
25:44many experts who can develop software for our people. And also because our people have less resources, less
25:52money to hire those highly paid expertise to build software for ourselves. So let's be more broad
26:00about information tools, okay? Like Excel is an information tool or a spreadsheet, a word template is also an
26:12information tool. It's larger than just the software. But anyways, my point is there is always a
26:20structural insufficiency of information tools supplied to the Global South. Global South, especially the Global
26:28people who are not well served for information tools. We have a lot of situations where we need information
26:37tools to help our people, but we don't have such tools because we don't have the resources. I think a
26:44great
26:45opportunity that AI brings is that now we have the opportunity, we have the possibility to build information
26:53tools for our own people. Let me give you an example. I recently worked with a TV station, Pan-Africanism
27:02TV
27:03in Ghana. You probably know them. They are a small TV station, much smaller than you are. They have about
27:1080
27:11people in total. Many of them are just volunteers. And they don't have the capacity to have world-class
27:21commentary for emerging news. They had to rely on some experts to write comments for them. So
27:32thanks to the development of AI, after 10 days training, the editors of PA TV was able to build such
27:41a,
27:41we call it agentic system. So it's an intelligent system composed with a few agents. So agents are
27:52independent sort of units can perform some particular task. For example, one agent is a deep researcher. It
28:01can go to the internet and collect any news, any sources about a news thread, things like that. So they
28:11created such a system. It has six or seven agents, and then those agents work on a large language model.
28:21Now they are able to, with any emergent news happens, they are able to write a world-class
28:29commentary. Of course, still they need to interview some experts to get a few sentences comments, a few
28:37very short but sharp opinions. Then they can write a world-class commentary in a few hours.
28:44I think that level of capability is the biggest opportunities that AI brings us. Think about
28:51PA TV, right? Who would build a software for them? Nobody. Nobody builds software for them because they
28:59don't have money. They barely have money for their own salaries. So the software companies would never
29:06go to them and try to build a software, a commentary, a news production system for them. But now, with
29:13the
29:13help of AI, they can build such a system for themselves. I think that kind of story in the grassroots
29:20organizations, media, think tanks, universities, and social movements, they should be able to build a lot
29:29of information tools for themselves. They don't need to rely on anyone, and nobody can cut their supply of
29:36information tools. I think that is an important opportunity.
29:39So what you're saying is that in these cases, used in the way that you're presenting it, AI could be
29:49useful in
29:50order to bridge certain gaps regarding investment, technical capacity, in order to do things that otherwise
30:00these grassroots organizations, projects could not be able to do because now they have this technology at their
30:07disposal. That is super interesting to keep in mind. Also, a little bit of optimism as well. We hear so
30:15many terrible things about what's coming. So it's good to have that in the horizon. Now, of course, everything
30:21that we have on our conversation, we could continue talking. I think we could do a podcast on every one
30:27of
30:27these items. But you touched on communication and the way this is being used. And there's a lot of
30:34of other risks that we are seeing and direct impacts that we are seeing right now of AI in communication
30:43of maybe other uses that are having more negative impacts. We have communication services using AI,
30:51and so reproducing Western stereotypes that are just repeating what main Western media is saying,
31:01because they are learning from those contents themselves. So that is making the battle for
31:09ideas from the global South so much harder. In that way, we are seeing AI also collaborate in the spread
31:20of
31:20misinformation, for example. Well, a lot of that has to do with political campaigns and how the far right has
31:27used it. So particularly in communication, because I know that you have this perspective of seeing
31:34everything that could be done. What do Global South communication outlets need to keep in mind in
31:42order to stay away from the main risks and also use AI to their advantage? This is a difficult one,
31:51because the media sphere is so complicated and it's so weedy. It's not a technological issue anymore,
32:08it's a political issue that the social media platforms are not being regulated.
32:17When they are being regulated, they're regulated as
32:21like you cannot talk about supporting Gaza and you cannot support Cuba. And then when other people
32:31are spreading fake news on those... This is pre-AI, right? This is before people use AI to generate those
32:42content. Look at Donald Trump. Donald Trump is making fake news every day and then manipulates the stock
32:50market, but nobody says anything to it. So it's already not a technical issue, but of course,
32:58being having the technology in hand makes them more capable of creating those fake news and disinformation
33:06and flood the internet with hatred and bias. I don't know the fundamental answer, right? You know, China
33:15is able to prevent itself from those misinformation by creating great firewall, by not allowing West
33:27so-called mainstream media and social media to propagate in the internet of China.
33:36I personally believe the regulation of the internet is part of the sovereignty of a country.
33:46It doesn't make sense to say that cyberspace is a free, open space, that there is no territory in the
33:54cyberspace because when you say that, it's basically saying as a national state, you should give up any
34:01regulation and give the power to Google and Facebook. So my opinion is national states should regulate
34:10internet space. I know it's a difficult task. There is a lot to do to achieve that, but that is,
34:17I think,
34:17the fundamental solution to that. Then back to Global South Media, I think a
34:27urgent task for us is we need to learn the technology and to use it, use it properly and use
34:35it wisely,
34:36because the problem you mentioned, almost all the mainstream large language models have
34:43a very clear pro-West bias, or be more specific, a pro-West liberal bias. When you talk to those
34:52models
34:52without any fine-tuning, you feel that you are talking to an American Democrat. But that can be,
35:00that political tendency can be adjusted, and it's quite easy to be adjusted, I would say.
35:07What you need to do is just set up your ideological and political framework for the AI to follow.
35:16For example, in my own case, I have a very thorough Marxist framework, because I'm a Marxist.
35:25So I tend to use Marxist theory to understand and to analyze the world. For example, I would always start
35:34from historical dialectics and materialism. So I created a seven-layer Marxist ideological framework,
35:46and I always tell, before any task, I always inject that prompt. So it's basically a prompt. I always
35:55inject that prompt to the AI that says, you are a Marxist theoretical analyst. You should think,
36:03follow this framework. And then here is a task. I found it's quite effective. It can suppress a lot of
36:13the pro-West ideological bias, and it fits well for my task. Of course, I'm not here promoting Marxism.
36:20I'm just saying there are options. People should learn how to build their own ideological framework and
36:26use that to adjust the behavior of AI. That is clear. And also, I was thinking regarding what you were
36:34saying about how sometimes this conversation is being had in terms of lack of regulation or the need
36:43for regulation in terms of how much a state is regulated. And there has been a lot of campaign from
36:49the West regarding what you were just saying, right? Internet is a free space. And a lot of things that
36:57we know are not such, like what you were saying right now regarding the shadow banning and the just
37:04neglecting of all the posts regarding Gaza, for example, the ongoing genocide and Cuba, of course,
37:11Venezuela. So we know that there is regulation, that it's not a matter of having or not certain
37:19regulation of what is happening on that cyberspace, but rather who is taking those decisions, who has the
37:27capacity to implement them. And somehow in that discussion, they are being able to impose a narrative
37:35where what we are discussing is not really what is happening. We are sometimes entertained in media
37:42discussion regarding the importance or not of regulation. And from a Western perspective, the
37:51regulation, before you talked about how strict China is regarding certain aspects of AI development,
37:58and sometimes Western media tries to present that in a bad light, as if it were an over-regulation.
38:05And presenting the lack of such regulation as a positive thing, we're seeing in the spaces that
38:12Western operation is also heavily regulated. It's just heavily regulated with other principles and other
38:20actors in space, sometimes prioritizing the monopolies and their needs and wants from this cyberspace.
38:29I recently had an article on this topic. So it's basically a theoretical debate
38:35with Antonio Legri and Hart. It's about how do you identify the roles, the players,
38:43in political dynamics, right? Antonio Legri insists that in order to fight against the empire,
38:54okay, we all agree that the empire now already has a good combination of the United States as the
39:04government. It's military, it's intelligence, now it's technology. But the approach to fight against the
39:12empire, Negri insists, should be autonomous collective of the multitude.
39:20But I argue, as Robert Cox said, there are at least three important roles in the political space.
39:32There is empire, there is civil society, and there is national states.
39:38We should not ignore national states as the most, so far the most powerful
39:44way of collecting and uniting people in the global south. It's an achievement of the
39:52national independence movement, right? So we should not ignore the existence and the power of national
40:01states. The civil society should find a way to have a unification, to have a united front
40:08with the national states, so that they can together resist the empire. So I think that is when people
40:19talk about countries, especially a few important countries, China, Venezuela, Cuba, Vietnam, DPRK, Iran,
40:30right? You know what I'm talking about. They try to figure those countries like an evil
40:39authoritarianism and authoritarianism and try to convince a civil society to get away from national
40:46states. But that will only weaken the power of resisting and that will make the empire more easy
40:55to penetrate. Jumping particularly from what you were saying regarding the importance of strengthening the
41:01national state capacity of resisting these impositions in terms of the AI development as it's being done,
41:12for example, in the West, particularly from the United States. Thinking about how this is happening
41:18in Latin America, in particular, being it's a territory with the historical dependency in terms of
41:28the US using the territory, the resources and understanding it as its own area of influence,
41:34I think that is pretty much determining what is happening right now and what will happen in the
41:39coming years. And AI will be one of the key matters of discussion in that sense. We've been talking about
41:46the possibilities for global South countries and everything that could be done. But we are acting now,
41:53we are actually living in a world in which the US model for AI is the closest one right now
42:02in terms of at
42:04least the will that they have to control this territory that is Latin America and the Caribbean. What should
42:12these national states be really looking for in terms of what are the true risks that this Latin American and
42:21Caribbean states are facing in this region and in this area? What should they be looking for?
42:29And I think that we know the answer because you talked about a unified front already. But can national states
42:35in Latin America tend up for themselves alone? Or is it necessary to think of some sort of coordination?
42:42That's a good question. I think the first thing we need to do is to have alternatives. I'm not even
42:52saying replacements, right? I'm saying alternatives. If we are relying on a handful of American companies,
43:02we know what is going to happen, right? They are not going to open their technology. So the countries will
43:09not eventually have control to the technology. When they decide to cut the supply, they can cut the supply.
43:18That already happened to a few countries. And they can control the price because you don't have
43:23alternative. So that means you will pay higher cost of it. They will extract your data. And as we know,
43:31data is a new oil of 21st century. It has economic value. But Google, I suppose,
43:38is not going to have a mutual agreement with the Brazilian government and to say,
43:44let's develop our data together. I don't see that is happening. So the first thing, I think,
43:50is to have alternatives, just to have competition. And I believe even liberal economists will agree with
43:57me that competition is good for a healthy market. So where does that competition come from?
44:06China, very clearly. And it's not from China as a country or as an alternative supplier of those
44:15technologies. It's from an open ecosystem. Because those American companies, what we see is they are
44:23offering a closed ecosystem. They try to bond you. They try to reference the users into their own
44:31ecosystem so that they can have higher profitability. The alternative is an open ecosystem. Latest larger
44:39language models become commodities. And the users can freely choose which model they use and they can
44:47build their own applications on top of it. So that's the first step. Then the second is, I think there
44:57is a lot
44:58lot to learn from the, please forgive my arrogance if it sounds like so, but I sometimes I feel myself
45:07swinging in the middle of being a Chinese and being a Latin American. So I sometimes speak like
45:14this. And I think there's a lot of things we can learn. So this is, this is the right place
45:19for you then?
45:22Yeah. I think there's a lot of things we can learn from the China's AI plus model. It shows
45:29the development of AI is not for the sake of AI itself. It's not for the stock market, the stock
45:39price
45:39of a few companies. If you look at the NASDAQ today, right, it's basically a few companies are leading
45:46the growth of the growth of the stock market. That is not what China is looking for. China,
45:52through AI plus strategy, China is trying to make AI a useful tool for every other sector and to increase
46:03the efficiency of every other sector and to help people from having to work on those boring and
46:10the dangerous and heavy work. So I think an AI plus strategy can also benefit many global South
46:19countries. You don't have to have world class, large language model development capacity. To be honest,
46:27that is difficult and that is expensive. But even without that, you can use the technology to create
46:35applications that benefits your own people. And then, of course, there are some security bottom lines
46:43you have to pay attention to, like continuancy of supply, the ownership of data and the right to control
46:52and the right to regulate and the right to decide which direction to develop. I think there is a tendency,
47:01overall there is a tendency to overestimate the value of foundational research and development and
47:11underestimate the value of applications. Building applications that the people need that can help
47:19the people is also important innovation. We should not underestimate that. We should not undermine
47:25the innovation of building useful tools for the people. So you can see my t-shirt. This is from
47:35Allende 1972. Allende said, we must create the technology proper to our own reality. He didn't say that
47:47as foundational research and development. Back then, when they tried to build an IT system in Chile,
47:55they didn't have their own computers. They bought two IBM computers to build the system.
48:00But Allende can see the value of building applications to serve the people.
48:05So he said, we must create the technology. He didn't say we must adopt the technology. Because
48:14to understand what the people need and build the applications, the information tools
48:21to fit the people's needs. It's also innovation. It's also creation of technology. So I think
48:28Global South countries should not undermine that. And we should bring the importance
48:36of making applications, making technology to support every other sector and to serve the people,
48:44to give a priority to that. We definitely hope so, too. And it's all about opening those
48:50new horizons, the possibility of thinking about this future in which new opportunities can come from
48:58the grassroots, including AI. And we hope that's the way we are heading. At least that's the hope.
49:05Thank you so much, Jeff, for joining us here in overlap time. Thank you very much. So that was one
49:11more
49:11episode of overlap conversations from across the world. In this case, going to the world of AI,
49:18we will meet again for now. See you next time.
49:35And this was another episode of Shining Hour, a show that opens a window to the present and
49:40future of the Asian giant. Hope you enjoyed and see you next time.
Comments