Stay informed with free updates
Simply register Artificial intelligence Myft Digest – delivered directly to your inbox.
Two topics dominated the annual meeting of the World Economic Forum in Davos last week: Donald Trump and artificial intelligence. Of the two, the latter was all the more interesting and almost more important. A lot of attention in the discussion was the task DeepseekThe surprising Chinese climb. Nevertheless, we have only learned that knowledge spreads: no country will monopolize these new technologies. This surprised the markets. With new technologies, such “surprises” are not surprising. But the big question does not change what all of us mean for all of us progress for the intelligence of the machine.
People are both social and intelligent. This combination is your “killer app”. It enabled them to dominate the planet. Human intelligence invented the all -purpose technologies that have shaped the world. From the taming of the fire to the creation of computers. But with computers who think that could change. Blaise PascalThe French mathematician and philosopher from the 17th century, said that “Man is only a reed, the weakest in nature, but he is a thinking reed.Is this uniqueness over now?
In Davos I took part in two fascinating discussions about the rewards and risks of progress in the AI. One was one an interview from Sir Demis HassabisCo -founder of Google Deepmind and joint recipient of the Nobel Prize for Chemistry, by Roula Khalaf, Editor of the FT. The other was one interview from Dario AmodeiFounder and CEO of Anthropic and author of Machines of loving gracefrom Zanny Minton BeddoesEditor of the economist.
The interview with Hassabis underlined recent progress in our ability to carry out scientific analyzes, especially in biology. Use more than 2 million researchers Alphabhe said the Deepmind program has developed. “We have all folded proteins known to science, every 200 mn. . . (T) The rule of thumb is that a doctoral student accepts its entire doctorate to find the structure of a protein. Therefore, 200 million doctoral thesis would have taken a billion years. And we only gave that for free in the world. “This is” science at digital speed “. The possibility that opened us in front of us has enormous acceleration of medical progress. In fact, we could have the next 50 to 100 years of normal progress in five to ten years.
On the whole, Amodei argued, we can “imagine”A country with genius in a data centerOne who may have made the Chinese even cheaper than before. But are they really genius? My test would be whether KI could produce Einstein’s general theory of relativity in view of the knowledge of all physics until 1906.
It seems plausible that the effects of such a problem -solving capacity, whether “awesome” or not, should be remarkable. Among other things, it could accelerate improvements in knowledge and productivity growth and the spread of prosperity. Both are desirable. In recent decades, the increase in “total factor productivity” – the best degree of technical progress – has been modest. In addition, large numbers still live in extreme poverty and, depressing, The progress has slowed down.
However, it is also obvious that accelerated progress can also lead to difficulties. The structure of the labor market could change massively, for example with a strong decline in demand for employees whose assets are trained, but largely routinely, intelligence. The forecasts of such effects vary. A 2023 Paper By Erik Brynjolfson and Gabriel Unger notes that the effects on productivity in the entire computer revolution, as it applies during the entire computer revolution, could be modest. This time, however, it may be different, with increasing productivity, but correspondingly great and disruptive economic and social changes. Depending on how society reacts, a successful AI can lead to “techno-foosalism” with even greater concentrations of prosperity. The invention of a large number of new treatments can significantly increase the costs of health care and the costs for coping with much prolonged life, even if they are healthier in balance. Are people ready to live next to their great -grandparents? Apparently good things could create real challenges.
In addition, the development of the planned AI poses great risks. How do you control your use by villain actors, including enemy states, terrorists and mass murderers? What moral judgments do you allow AI in warfare? How do you control the use of AI in surveillance? Will “Big Brother” be watching us more and more? What do we do again against the production of counterfeits and fake news? How does freedom survive all of these threats?
Hassabis is clear that we need effective global limits to use AI. Will China and the United States work together in a time of broken international cooperation and contempt for the idea of a “rules on international order” to protect AI? It appears unlikely, not least because they have different views of how such technologies should be used.
Already in 2015 I wrote a generally skeptical Article On the (modest) likely effects on the productivity of new technologies. The next few years could finally prove the opposite. But I also noticed that if we approached instead “The singularity” – Artificial intelligence that exceeds all human intelligence – everything has to change.
One of the great ideas with Frank Herbert dune Series is that in the distant past (our future) humanity has led a successful jihad against machines that think. After that, people had to be superhuman. A leading character explained That “people had made these machines to be used to use our feeling of beauty, our necessary selfom from which we make living judgments. Of course, the machines were destroyed. “
This concern could prove to be wise. But I am realistic: AI is made of Pandora’s box.
