Getty images
The Chinese application Deepseek has shaken the world of AI
As I write, there are many powerful people around the world to prepare to go to Paris – and a feeling that many of them hold their breath.
Monday, in the sumptuous context of the Grand Palais of 125 years of the city, representatives of 80 countries, including world leaders, technological bosses, academics and other experts, will meet for a two -day world summit to discuss Current progress and future objectives, for rapid and extremely disruptive technology which is artificial intelligence.
This is perhaps what is on the official agenda of the summit of the action of artificial intelligence, but there is something else that breathes fire in this particular speech store: Deepseek.
A fan of horse races told me once the day before a big race, everyone is a winner.
And with China considerably blowing AI competition in a new direction with Deepseek, its super efficient and super-viral AI assistant, there is suddenly a feeling in front of the summit that the pole position occupied by the American AI sector, despite its vast wealth and IA infrastructure, may not be so out of reach after all.
Professor Gina Neff, from the Minderoo Center for Technology and Democracy at the University of Cambridge, says that there is currently “a void for world leadership on AI”.
Professor Dame Wendy Hall, from the University of Southampton, agrees. “Deepseek has made everyone realize that China is a force with which it is necessary to count,” explains the computer scientist.
“We don’t have to accept what big companies on the west coast say. We need a global dialogue.”
On this front, the time of the summit could not be better.
Europe also spies on an opportunity to make a new offer for the AI crown. One of the officials of French president Emmanuel Macron described the summit to journalists as a “alarm clock” for France and Europe, adding that the block must not leave the AI revolution “pass it”.
Other countries also recognize a potential change in the power of AI in the air. India, Prime Minister Narendra Modi, confirmed his attendance at this summit – not having come to previous gatherings.
The United States send a serious firepower as a defensive signal that is its own, including vice-president JD Vance, the CEO of Openai Sam Altman and Sundar Pichai of Google.
Elon Musk is notably absent from the official list of guests, but he will undoubtedly have something to say about it, whether in person or not.
British Prime Minister Kier Starmer would also be away.
Getty images
US vice-president JD VANCE will direct the American delegation
There were two previous summits, the first held in the United Kingdom and the second in South Korea. Many things have happened since the IA world met for the first time, on the steps of the historic manor of Bletchley House, in November 2023, and promised to try to maximize the advantages of the While minimizing risks. On the one hand, half of the world’s population has been in the polls since then.
In Bletchley Park, Chinese Vice Minister of Science and Technology, Wu Zhaohii participated. But there were whispers that he was kept on his arms due to national security.
In Paris, in comparison, I expect China to be guests of honor. The country would have sent one of its senior leaders, Ding Xuexiang, a close ally of President Xi Jinping. There is also a buzz to find out if the creator of Deepseek Liang Wenfeng will join him.
I asked Chatgpt to list some of the notable stages of AI from the Southern South Korea in May 2024. Deepseek did not cut.
Of course, there is much more in AI than generative AI like Deepseek and Chatgpt – tools that create content such as text, images, video. This could well be the most accessible as consumers. But there are also AI tools that drop the symptoms of the disease, model climate change solutions, develop new formulas for drugs – and all of this will be in the range of Paris.
In addition, David V Goliath’s account around which Deepseek’s depth is blocked deserves a more in -depth examination. The boss of the company of IA Anthropic, Dario Amodei, wrote a convincing blog to find out if Deepseek was really built at a fraction of the cost of his American competitors.
We know that it was built on their shoulders: using a number of NVIDIA chips (probably older, due to American sanctions) and an OpenSource IA architecture developed by Meta. In addition, Openai complained that competitors use his work in order to advance their own (contacts in the creative industries were amused to the irony of this, since Openai products will readily produce production “in style of “individual human creators).
But nevertheless Deepseek managed to shake up the AI sector in a way that even the AI itself could have predicted. And that suffered a lot of money on the value of some of the greatest players in the process. It will almost certainly be a huge subject of conversation all around these Parisian conference rooms.
Getty images
The summit takes place at the Grand Palais Building in the center of Paris
There is another theme that crosses the summit of AI which will be worth keeping an eye on.
The first summit had the word “security” in its title. Some estimated that the event pushed the narrative that is too hard and terrified people with dark speeches of existential threats.
But it did not fall completely from the agenda.
As a subject, AI security is a rather wide church. It can be linked to a certain number of risks: the generation and propagation of disinformation, demonstrations of biases and discrimination against individuals or races, the development in progress by several countries of arms controlled by the ‘AI, AI potential to create unstoppable computer viruses.
Professor Geoffrey Hinton, often described as one of the sponsors of AI, says them as “short -term risks”. They could be discussed in Paris, but he argued last week on the BBC Radio 4 bbc program that they are unlikely to obtain a strong long -term international collaboration.
The great scenario which, according to him, will really bring together everyone is the prospect that AI becomes smarter than humans – and want to take control.
“No one wants the AI to take over from people,” he said. “The Chinese preferred by far that the Chinese Communist Party directed the show that AI.”
Professor Hinton compared this possibility up to the Cold War, when the United States and Russia “succeed roughly” in collaborating in order to prevent the world nuclear war.
“There is no hope of stopping (AI development),” he said. “What we have to do is try to develop it safely.”
Professor Max Tegmark, founder of the Future of Life Institute, also shares a striking warning. “Either we develop an incredible AI that helps humans, or an uncontrollable AI that replaces humans,” he said.
“We are unfortunately closer to the construction of AI than to understand how to control it.”
Professor Tegmark hopes that the summit will pressure for the connection of security standards “as we have done in all other critical industries”.
Learn more about global trade and technological stories
Source link
