Oppenheimer and the AI ​​Prometheus

Are we being overly scared with the threats to humanity posed by Artificial Intelligence (AI) after the emergence, in November 2022, of ChatGPT and other generative AIs, and what’s to come? The curious thing is that those who spread fear the most are those responsible for having developed it, the AI ​​Prometheans, who are now asking to be regulated. Like Robert Oppenheimer, head of the Los Alamos National Laboratory, coordinator of the Manhattan Project that built the atomic bomb, he came to have doubts – some versions speak of “regret” – about its use in Hiroshima and Nagasaki because he considered that the Japanese were already defeated. . “We knew that the world would not be the same. A few laughed, others cried, most were silent, ”he would later say. Oppenheimer subsequently opposed the development of the 1,000 times more powerful hydrogen bomb (but did not resign from his position as adviser to the commission dealing with the issue). He requested, at least, international control over the new weapon of a destructive capacity without equal in the history of humanity. The comparison is relevant because there are several people responsible for AI who equate its dangers with those of nuclear weapons (again on the table with the war in Ukraine).

Christopher Nolan’s movie about Oppenheimer is about to be released. One of its main bases is the book by Kai Bird and Martin J. Sherwin, American Prometheus: The Triumph and Tragedy of J. Robert Oppenheimer, although it will surely contain controversial interpretations. President Harry Truman called the physicist a “whiny scientist.” Nolan, in a recent interview, established some parallels, but also differences, between the bomb and the AI. “It’s the best analogy – and that’s why I used it on Tenet – for the dangers of mindlessly unleashing a new technology on the world.” The philosopher Ortega y Gasset, after World War II and the bomb, already wrote that “the ability to build a world is inseparable from the ability to destroy it” and that “the prodigious advance” of technology “has given rise to inventions in which man, for the first time, is terrified by his own creation”.

Well, there are beginning to be many scientists, and businessmen, whining about the new AI that, although it opens up immense new possibilities, could eventually control humans, and treat us like pets if we do not exercise control over it. At the moment, it poses problems of employment and work; of competition, with the possibility that, once again, a few large companies will control AI in the world; environment; and on a slightly further horizon, but not so much, of its military applications and decisions about life and death in the heat of a conflict. In fact, an arms control expert like the centenarian Henry Kissinger has called for international agreements, including a military AI control agency, and the stabilization of the race to dominate it.

“Control and regulation is difficult when you start not knowing what is happening, how the AI ​​of advanced neural networks processes your information”

The AI ​​Prometheans – Elon Musk, Bill Gates, Geoffrey Hinton and Mo Gawdet (the latter two left Google after leading or working on its most advanced technology), Sam Altman, president of OpenAI (creator of ChatGPT) – call for regulation of the great projects in this field, accompanied by many others of the thousands who have signed the public appeal from the Future of Life Institute for a six-month moratorium on new “giant” experiments with AI. They see a danger if this technology runs amok, even though many of them are investing heavily in it. Even China warns of potential catastrophes for the misuse of AI. They all know that control and regulation is difficult, especially when you start not knowing what’s going on, how AI from advanced neural networks and other cutting-edge techniques are processing your information. Statements of principles or proposals in progress, such as the one that is shaping the European Union, they are very good, but they forget about the technical difficulties, of engineering, of taking those terms to the coding and operation of the AI. In addition, they leave in a very secondary place the impact of the enormous energy expenditure –and therefore even greenhouse gas emissions as long as all this energy is not clean– of the new AI, which, however, is also essential when it comes to design measures for the recovery of the environment.

Advertising

Regulation yes, these Prometheus point out. But, deep down, it is to protect companies. Big technology companies are asking for regulation to protect their interests, their patents, to avoid payments for copyrights on the gigantic information – texts, images, sounds, including music – with which they feed their systems. The president of the Spanish government, Pedro Sánchez, may have asked the United States, Joe Biden, for transatlantic collaboration in the development of artificial intelligence –“Europe has to be a voice that is heard clearly”, he said in Washington–, but the reality today is that the main AI companies in the world are from the US and China (even Google acquired the British DeepMind). Excluding China, of the 15 older AI companies in the world, led by Microsoft and Google, 13 are from the US, one from Israel and another from the US, UK and Singapore. None from the EU, which has plans but is always lagging behind. Europe is ahead in regulation, but behind in this technology that is changing the world. The stakes are high: the control by a few companies of the two superpowers over the citizens and the European economy. Big AI will be worse, economists say Daron Acemoglu and Simon Johnson, authors of the excellent Power and Progress: Our 1,000 Year Struggle Over Technology and Prosperity. In addition, companies such as Alphabet (Google), also Amazon, among others, are warning his employees to limit their use of generative AI, including his own like Google’s Bard, to prevent leaks of his technology.

We must strive to ensure that AI serves the common good, the good of all. These days the third AI for Good Summit. It is worth remembering, with the historical review that Acemoglu and Johnson do, that there is no automatism that makes the new technologies bring prosperity for all, and the AI ​​Prometheuses say little about this. “Whether they do or not is an economic, social and political decision.” In Greek mythology, Prometheus was a Titan who stole fire from the gods to give it to humans, for which he was punished by Zeus for eternity.