- 19 minutes ago
In this India Today Programme Video, anchor Marya Shakil takes up the intensifying global race to develop artificial intelligence and the growing debate over whether humans can keep pace with the technology they are creating. With tech leaders including Anthropic CEO Dario Amodei, OpenAI’s Sam Altman, Elon Musk and Microsoft CEO Satya Nadella warning about AI development outpacing safety guardrails, the question of how far the AI race should go is becoming increasingly urgent. US President Donald Trump, however, has rejected calls for slowing down, arguing that the United States cannot afford to fall behind China.
Category
🗞
NewsTranscript
00:00Good evening, you're watching NewsTrack with me, Maria Shakil.
00:03Our top focus tonight, AI versus humanity, control war.
00:07As the race to build powerful artificial intelligence accelerates,
00:13a far more fundamental question is emerging.
00:16What happens if AI gets there before we humans are ready?
00:20From Washington to Beijing, AI is no longer just a tech story.
00:25It is about power, national security, and who gets to shape the future.
00:30But here is dangerous paradox.
00:34The dangerous paradox.
00:37Even the people building some of the world's most advanced AI systems
00:41are now raising the alarm,
00:42warning that the tech may be advancing faster than the safeguards
00:46designed to keep it under human control.
00:50The world knows AI could be risky,
00:52and yet no major power wants to be the one to slow down first.
00:57Because the world may now fear falling behind the AI race
01:01almost as much as it fears AI itself.
01:05And that is where the biggest question come to the forefront.
01:09If AI is dangerous, why are we racing to build it?
01:13Before I bring in the guests, here's a report.
01:34The world wants to win the artificial intelligence race.
01:38But suddenly the biggest question is not just who gets there first,
01:43but what happens if AI gets there before we humans are ready.
01:49At the BRICS summit, Prime Minister Narendra Modi warned about the weaponization of technology.
01:55But Chinese President Xi Jinping pushed for an open-source AI ecosystem within BRICS,
02:01a model of cooperation that puts China at the center of the global AI conversation.
02:07But now that conversation is turning into a confrontation.
02:12After former Enthropic researcher Jacob resigned and warned that AI could kill us all by the end of the decade.
02:21Enthropic CEO Dario Emodei has called for the world's most advanced AI companies
02:26to slow the pace of developing warning that AI is improving so quickly
02:31that safety and human control could fall behind.
02:35Open AI's Sam Altman, too, has agreed.
02:39Tesla boss Elon Musk has also said Dario is right.
02:44Even Microsoft's Satya Nadella has now added his voice,
02:48saying any pursuit of superintelligence must benefit humanity and remain under human control.
02:56But Beijing is pushing back.
02:58China has rejected cause to curb its AI development,
03:02saying countries should cooperate rather than engage in confrontation
03:06of what it calls malicious competition.
03:12The development of artificial intelligence concerns the common well-being of all humanity.
03:18Spreading various threat narratives and engaging in confrontation and malicious competition
03:22will only disrupt the global AI governance process and serve the interests of no one.
03:31And China's state-backed Global Times has even accused Emodei of using a Cold War playbook
03:38to contain China's AI ambitions.
03:40There is a reason this is becoming so explosive.
03:44Emodei himself has warned that a Chinese lead in AI could pose a grave danger to the US
03:50while calling for tighter restrictions on advanced AI chips going to China,
03:56a charge that Beijing has denied.
04:13So the argument is no longer simply AI safety versus AI innovation.
04:18It is becoming AI safety versus the fear of losing the AI race.
04:23And who else is firmly on the side of that race?
04:27US President Donald Trump.
04:29He has dismissed cause to slow AI development, citing rivalry with China.
04:35We're leading China in AI.
04:37We're the most sophisticated country in the world.
04:40And frankly, I want to keep it that way because whoever wins, AI wins.
04:44And we can put guardrails, we can do this and that.
04:47But I think you have a lot of negative voices that are bringing it up that shouldn't be bringing it
04:52up.
04:53And they're bringing up things that won't happen.
04:55And that indeed puts the coming Trump-Xi talks in a very different light.
05:00Because AI is now sitting at the intersection of tech, national security and global power.
05:06The irony is striking.
05:08The US and China are preparing to talk about AI safety.
05:12But both are also racing to make sure the other does not get ahead.
05:18While governments fight over who controls AI, its builders face a deeper question.
05:26Can humans remain in control of what they are creating?
05:30And now, people are asking them something even simpler.
05:33If it is dangerous, why build it?
05:37If AI capabilities race ahead of safeguards, the biggest risk may not be who wins the AI race,
05:44but whether anyone is ready for what comes after the finish line.
05:49Bureau Report, India Today.
06:00And breaking news coming in now, even as tech czars issue warnings and call for a slowdown,
06:06US President Donald Trump has said that the only reason the AI and data center outburst is happening
06:13is because the United States is leading by a lot every other country.
06:19He also went on to say that the only control or guardrails that AI needs is a strong and smart,
06:26high IQ president and the USA has that in spades.
06:30The Trump administration has stopped AI people from doing bad or potentially bad things,
06:36like Dario, who is now pretending to be a perfect little angel, and we will continue to do so.
06:41We already have tremendous criminal and regulatory power over these companies.
06:46There is a sick conspiracy going on against AI and data centers,
06:49and the only one that is happy about it is in China.
06:55So those are the words coming in of US President Donald Trump.
06:59Let me bring in Gary Marcus, cognitive scientist, and he's joining me from Vancouver, Canada.
07:07Gary on X says that he is the original OG Gen AI skeptic.
07:15So, Dr. Marcus, welcome to this broadcast.
07:20You have said, Dr. Marcus, that it is preposterous that humans will be extinct in five years
07:26and equally preposterous that everything is going to be fine.
07:30The two most dominant narratives are entirely wrong.
07:34Those are your words.
07:35If both extremes are wrong, what exactly do you place Jacob, where do you, I should say,
07:43where do you place Jacob Coxon's warning that the anthropic researcher
07:49who quit saying that labs are, in fact, gambling with our lives?
07:55I think the specifics of what he has said are preposterous,
07:59and he has kind of, like, doubled down in a series of interviews.
08:03So he is very much on the side of, like, we might all die in five years,
08:07and I think that's ridiculous.
08:09But I think that there are serious warnings.
08:12People in the field talk about P-Doom, the probability of extinction.
08:15I think that's pretty close to zero.
08:18You know, if the robots come for us, we will fight back.
08:22We're geographically diverse.
08:24We're genetically diverse.
08:25AI is not going to wipe out the entire population.
08:27But there's also what I call P-catastrophe, the chance that something serious might happen.
08:33If something serious means, for example, electrical grids being knocked out by cybercrime,
08:39maybe by malicious individuals using these technologies or from disinformation causing accidental wars,
08:46I think that that's very serious.
08:48I think that we probably will have some very serious side effects.
08:53We're not going to literally have extinction.
08:54People don't need to, I think, have nightmares about that.
08:57I mean, eventually, you know, the sun will explode or something in the very long term.
09:01But in the near term, we don't have to worry about those things.
09:04But we do have to worry, excuse me, my voice is giving out because I've done so many interviews.
09:10We do have to worry about the consequences of insecure AI, basically.
09:17Another way I would put it is, like, a lot of AI is fine.
09:19Like, GPS navigation systems are not going to kill anybody.
09:23You know, people basically know how to make that software.
09:26But we have this new kind of software called AI Agents, where you take large language models that are not
09:32reliable,
09:33you hook up a lot of them, and you connect them to the Internet.
09:36And sometimes they do bad things.
09:38The OpenAI Hugging Face Incident was an example of that,
09:41where a set of AI agents couldn't follow basic rules like don't hack into other people's systems.
09:48And that inability to follow rules is where the danger really comes from.
09:51It's not because AI is too smart.
09:53It's because current AI is not that smart.
09:56It can't follow rules.
09:57That means it can be exploited by bad actors.
10:00And it means that it might just misunderstand instruction and do something bad.
10:06Not extinction level, but serious.
10:09So, like, Jacob Coxon, I think, is way overstating the case.
10:13But there is something to be worried about.
10:15We can't just sit here and have no regulation and hope for the best,
10:19as some world leaders may have suggested.
10:21Okay, Dr. Marcus, I'm going to move away from what Jacob said to BRICS summit here in New Delhi.
10:29Prime Minister Modi told BRICS that weaponization of technology can hinder shared progress,
10:36calling for inclusive access.
10:38Is the danger you see the same one Prime Minister Modi is naming nations using tech access as leverage?
10:46Or is the more urgent weaponization happening elsewhere, perhaps in how AI capability itself is being marketed?
10:58I mean, there's a lot of different things going on.
11:01It's hard to say, like, which is most important.
11:03There is, though it's not really what I was talking about, issues about equitable access to these technologies,
11:10about how the countries other than U.S. and China are going to have access and what that's going to
11:16look like.
11:17So there's a whole set of issues about economic distribution and things like that.
11:23There's another set of issues, which is really one that I'm talking about,
11:26which is that current AI is not very well controlled, or people use the technical term aligned.
11:31And that can mean it does things we don't want it to do, or that bad actors can exploit it
11:36to do bad things.
11:37And that's a different set of risks than the economic set of risks.
11:42And there's still other risks, like non-consensual deepfake porn is yet a third category of risks
11:49that is, you know, starting to cause a lot of problems in high schools and so forth.
11:53The thing about AI is it's not one thing, but many.
11:57There's many different underlying technologies with different capabilities.
12:00There's all kinds of social and political things that are factoring in.
12:03So we can't just say, like, this is the one thing you should worry about.
12:07The truth is it's like a hydra with many different heads.
12:10There are also many advantages, but there are many disadvantages we need to worry and to rein in.
12:15Then are you worried more about disadvantages than advantages,
12:19given the pace of AI and the concerns being raised now by the big, you know, the tech giants themselves?
12:30I'm pretty worried at this point.
12:31I think I'm long-term positive in the sense that I think that there is a possible path to an
12:38AI
12:38that is really good for society.
12:40But short-term, I'm worried.
12:42I think the current technology is unreliable.
12:44And I think that unreliability causes problems.
12:47And I think that the companies that are running it are mostly pretty greedy
12:51and not really mindful of the consequences.
12:53And I think the governments are mostly letting those greedy companies do what they want
12:57at the expense of humanity.
12:58So right now, I think AI is probably a net negative for society.
13:02I don't think it has to be that way, but I think that's kind of where we are right now.
13:06And we need to up our regulation game.
13:08And we need to have better technical tools as well.
13:12So what can be these regulation and technical tools?
13:15What's responsible AI then?
13:18Or should I say, how do you make AI responsible?
13:24On the governance and regulatory side, I think the number one thing that we need
13:28is that every country needs to do something like what the U.S. does for drugs with the FDA,
13:33which is to say, we require a cost-benefit analysis.
13:37You have to show that the value of your drug offsets the consequences, the risks.
13:43Lots of drugs have side effects.
13:45That doesn't mean we don't put them on the market.
13:46But you have to make the case that the benefits outweigh the risks.
13:50And you have to have independent scientists weigh in on that.
13:53You have to make an argument that what you have is good.
13:57And they may send you back to the drawing board or ask you to do more experiments
14:00or come up with a way of dealing with something.
14:02We need that for AI.
14:04That's the number one thing that we need on the regulatory side
14:07is some kind of screening procedure.
14:09It needs to be transparent.
14:10It needs to have independent scientists in the loop.
14:14Secondly, on the technical side, we need AI that can follow instructions.
14:18That might sound trivial, but it's actually pretty hard given the starting point we have now.
14:24So the starting point that we have right now is these large language models
14:28that try to predict essentially what a person would say in a given context.
14:32And it does that by looking at the statistics of how people talk
14:36across a whole lot of different kinds of text and so forth.
14:39That system is not built from the ground up to follow instructions.
14:43It often follows instructions, but it's not guaranteed to do so.
14:46And that's a real problem.
14:48And I think we need to go back to the drawing board for better technology.
14:52So, Professor Marcus, China here is pushing open source AI models
14:56to the global south.
14:57And BRICS itself has just unveiled an open source AI ecosystem
15:01framed as tech equity.
15:04Is there a genuine or do you see this as genuine democratization
15:09or a subtler form of the same weaponization
15:13that Prime Minister Modi has warned against?
15:20It's probably both.
15:21You know, there's some value in having open source AI for the world.
15:25And it's China pushing its particular advantage.
15:28They're not as good on the closed source models.
15:30They're better on the open source models.
15:32And they want, you know, what we call vendor lock-in.
15:34They want people to use their models.
15:37So, it's a bit of both.
15:38It's complicated.
15:39Okay.
15:40So, the U.S.-China chip war is fought over
15:42who controls the physical layer AI runs on.
15:46If scaling is really a dead end, as you argue,
15:50has the entire chip war and China's open source counter push
15:55been fought over the wrong resource then?
16:01I think that people have been fighting over the wrong thing.
16:04So, right now, what matters maybe is who has more chips.
16:09But that's not going to last.
16:11What's really going to matter in the long run is who builds the better AI.
16:14And I think there's actually much more room to improve AI than people realize.
16:18We still have massive problems with planning,
16:21massive problems with these systems hallucinating,
16:24with reliability and so forth.
16:26I think that the real advantage comes from solving those technical problems.
16:29On the chip side, I think that China is going to catch up.
16:33The price of chips is going to go down.
16:35And people are actually going to lose a lot of money on chips when they depreciate
16:38and you have, you know, more chips than you really need.
16:41The other thing is that people are going to make more and more efficient systems.
16:46They're optimized to use fewer and fewer chips to run locally rather than on the cloud.
16:51And so, we're going to wind up with all these data centers we don't really need.
16:54And the advantage is not going to be in the data centers, I think,
16:58but in developing a better technology.
16:59But everybody is wedded to large language models right now
17:02and maybe not thinking outside the box enough.
17:05Okay.
17:05Professor Marcus, can a country then build real autonomy in AI
17:09by adopting China's open source models?
17:12Or does that just trade American dependency for Chinese dependency?
17:21This is, again, a complicated question.
17:24I don't know that any country can build autonomy at this moment
17:30except by building a better and maybe more creative technology.
17:36There's going to be some dependency right now.
17:38I think there's a lot of talk about building sovereign AI
17:41and I think more and more countries are going to do that
17:43because they don't want those dependencies.
17:46The positive thing for countries that might want to do that
17:50is that the general techniques that people are using are pretty well known
17:54and so it is possible for other countries to build their own.
17:58And as efficiencies come in place,
18:01I think the advantage of having giant clusters of GPUs will diminish.
18:05You won't need quite as many in order to do most of what you want to do.
18:10I have two more questions here, Dr. Marcus.
18:12After Jacob Coxon was, we saw Dario Amadei and also Sam Altman
18:18both spoke about slow down,
18:21that how there is a need to slow down the pace of growth.
18:24Where is the future then according to you?
18:28I mean, it's hard to predict the political future.
18:31It's kind of changing literally by the hour.
18:34One minute, Dario says, we should have a pause and Sam says, great.
18:38And then five hours later, Sam posts something on Twitter saying,
18:41well, we're not really stopping.
18:42We're just trying to do things a little bit more carefully or something like that.
18:47So there's a lot of back and forth about what these guys mean.
18:50There's back and forth about what the U.S. government wants to do.
18:54One minute they're saying they don't want to regulate anything.
18:56The next minute they say, we've got regulation
18:58and that regulation is something secret that we don't want to tell anybody about.
19:03And so the kind of political winds on this are changing constantly
19:07and I think are pretty unpredictable.
19:11In the United States, there's a lot of pressure against data centers.
19:15A lot of people are now concerned about so-called runaway AI.
19:19You know, Trump could change his mind on all of this on a dime
19:22and then we'd be in a different place.
19:24So I feel comfortable making technical predictions
19:27about what particular technologies can do.
19:29But on the political side right now,
19:31I don't think it's ever been more unpredictable than it is right now.
19:34And I don't think it's ever been more consequential either.
19:36Like, for example, whether or not China and the U.S.
19:39come to any deal about slowing down would have profound consequence.
19:42Maybe they'll do that.
19:44They probably won't.
19:45There's a good argument that they should.
19:47These decisions that she and Trump are going to make
19:51in the next several weeks are super important,
19:54but very hard to predict.
19:56Okay, so let's try and sum this up,
19:58sum this conversation up, Dr. Gary Marcus.
20:02You're saying that it's difficult to predict on the political side.
20:05What would you want the countries to do then?
20:11I think that the best thing that we could do
20:13is to stop thinking about AI as a zero-sum thing
20:16and start to think about AI as how we could collaborate
20:20and make the world a better place if we work together.
20:22That's not really how people are thinking about it now.
20:25They're all thinking about it as a race,
20:26but that race has not really helped anybody but NVIDIA.
20:29It hasn't really helped individual people.
20:31It's put economies like the U.S. perhaps in jeopardy.
20:34It's raised the level of tension in the world.
20:37Maybe we need a different paradigm that's much more cooperative.
20:40All right, Dr. Gary Marcus.
20:41Pleasure having you and getting your thoughts
20:44on what is appearing to be a very, very fast-changing world,
20:49even in the space of AI.
20:50Joining me now is another newsmaker, Dr. Arvind Gupta.
20:54He's the head of Digital India Foundation.
20:56Dr. Gupta, the Prime Minister has flagged
20:59the weaponization of technology as a global threat.
21:02And in the same forum, Xi Jinping offers BRICS Nation's open-source AI
21:07as an olive branch.
21:09So, is China's offer a genuine course correction
21:12or a calculated move to set the terms of global AI
21:16while the world debates the danger?
21:22Okay, Maria.
21:23Depending on the amount of time we have,
21:25thank you for, first of all, having me on this show.
21:28But the question that you asked,
21:31Prime Minister Modi, as always,
21:32whether it was the G20 summit
21:35or the global AI impact summit in India,
21:38or now the BRICS has maintained the non-weaponization of AI
21:44and should be a core fundamental principle
21:47as we progress on this frontier journey of AI development.
21:51But, you know, what we have fundamentally believed
21:54and why the PM talks about UPI and the Aadhaar
21:58and the vast digital public infrastructure that India has built
22:02is because the models are very different.
22:04And that is a fundamental thing that we need to understand.
22:06The Chinese model has always been technology for surveillance.
22:12The American model has been technology for capitalism.
22:15The Indian model, which the PM really champions,
22:18is the technology for development model,
22:20the humanity model.
22:21Now, when you take this into context
22:25and put AI also in this three different paradigms of technology,
22:30the Chinese are, you know, offering this olive branch
22:34or whatever the word we want to use,
22:37I think they just lack the trust.
22:41We, for example, in India don't use any Chinese technology
22:44because they were found to be using our data
22:48and their platforms against our own national interests.
22:53So, this Chinese offer of open sourcing,
22:59there is nothing called free in technology.
23:01Please keep this in mind.
23:02I mean, this open source models exist.
23:04You can use them.
23:05But because the Chinese system lacks the trust,
23:08I think this offer is, you know, it's not a great offer.
23:14There's nothing great that the Chinese premier has said.
23:17I would say that the Prime Minister Modi's statement
23:22and his request for non-weaponization,
23:25whether it's of critical minerals,
23:26which are critical to the hardware and the AI infrastructure,
23:29or to the big tech during the AI summit in India in February,
23:35is the way to go.
23:37And suddenly, what happened in February, six months now,
23:41right now, what we are seeing is the big tech is also realizing,
23:43suddenly, that there is a need for them to also agree
23:49on this non-weaponization.
23:51And if you remember, in February,
23:53open AI and Anthropic didn't even want to shake hands.
23:56So, you know, the world, six months is a long time
23:59in the AI world and it has come there.
24:01So, what we, you know, professed in February,
24:04the world is realizing it was more visionary
24:07than they could agree or form.
24:11Dr. Advin Gupta, even as, as you said rightly,
24:14even as world leaders, one of AI's dangers,
24:17here is President Donald Trump trying to downplay the risks.
24:22On the other hand, you have Silicon Valley CEOs
24:24quietly talking of slowing down.
24:26So, is Washington out of step with the rest of the world
24:30or is this more of a big techs just saying it,
24:36slow down, talk just to cover that they are falling behind China?
24:43See, the American models have a very big lead right now.
24:48And the American AI ecosystem is far ahead of China.
24:52The clear, you know, there is a clear race going on
24:56in the use of AI for, for economic development
25:00and just superiority, human superiority.
25:03You know, it's, it's a big race.
25:05Nobody's going to let go of this.
25:09While Donald Trump is, is making these statements,
25:13I think he, they also have a vested interest.
25:15They have this large Stargate project
25:16where they've committed investments
25:18of half a trillion dollar, 500 billion US.
25:21So, there's a lot of money at play.
25:23But eventually, Silicon Valley is realizing
25:27that there is a, there is dangers in this.
25:31Now, Silicon Valley is no,
25:35it's the most ironical people talking about dangers in this
25:39because they developed the technology
25:40without the seatbelts.
25:42And now you have cars running across these highways
25:44at such large speeds
25:46that the seatbelts have become necessary.
25:49And when we were saying, put the seatbelts
25:51and invent the seatbelts
25:52and invent, make sure the seatbelts are on
25:56in an analogy to the,
25:58to high-speed car driving on highways,
26:00they'd never listen to it.
26:02So, I think there is,
26:03there is,
26:04the Silicon Valley is at least talking a little bit,
26:08which is good.
26:10President Trump saying there is no danger of weaponization.
26:13He's protecting his national interest,
26:15which is AI superiority
26:17that the America already has.
26:19By doing anything else,
26:21they're going to,
26:21they're going to probably slow down a little bit.
26:24So, I think it's,
26:25these are all contravening forces at work.
26:30Clarity is not there.
26:32And I don't think so the air race is going to stop.
26:34It's, it's, it's progressing.
26:37A few, few people are disgruntled.
26:39People are leaving.
26:40And my question always is,
26:42if you, you know,
26:43you've been doing this for the last couple of years,
26:45why didn't you leave then?
26:46You know, this is not suddenly a wake-up call
26:49that they are getting.
26:50So, it is,
26:51this is,
26:52this is a very, very real danger that we face.
26:57But also, on the other hand,
26:58if you build the guardrails,
27:00make this responsible and safe use of AI
27:02as a guiding principle,
27:04which is what India's model has been,
27:07it works for everybody.
27:08It works for humanity.
27:09It works for development.
27:11All right, Dr. Adwin Gupta, pleasure.
27:13Thank you so much for your time.
27:14And today is Ganesh Chaturthi.
27:16Prime Minister Narendra Modi
27:17attended Ganesh Chaturthi festivities
27:20and performed in Aarti
27:21at Union Minister Piyush Goyal's residence today.
27:24I'm leaving you with these visuals.
27:25Thanks so much for watching.
Comments