ChatGPT, the new artificial intelligence (AI) program, has raised alarms that were not heard when similar technologies started to be applied to good advantage in fields such as machine translation. Clearly, ChatGPT threatens to produce kind of elaborated texts. This has touched the sensibilities of journalists, who are more likely than translator’s associations to put their fears on the agenda.
The press enjoys replicating catastrophist columns that confirm their fears, such as the one by Yuval Harari in The New York Times on March 24, 2023. His sentence is apocalyptic. “Artificial Intelligence could rapidly eat the whole of human culture — everything we have produced over thousands of years — digest it and begin to gush out a flood of new cultural artifacts. Not just school essays but also political speeches, ideological manifestos, holy books for new cults,” he said.
For a long time now, the press position vis-à-vis technology has been quite similar to that of politics: it embraces the worst predictions even before it has clear evidence.
In the same newspaper, but with less repercussion, Noam Chomsky, together with two specialists on March 8, 2023, affirmed exactly the opposite. From their solid knowledge of generative language, they expressed: “The human mind is not, like ChatGPT and its ilk, a lumbering statistical engine for pattern matching, gorging on hundreds of terabytes of data and extrapolating the most likely conversational response or most probable answer to a scientific question. On the contrary, the human mind is a surprisingly efficient and even elegant system that operates with small amounts of information.”
Before and After
That creative possibility is what artificial intelligence does not emulate. That a certain sector of journalism feels threatened by it tells more about the automatization of the work it executes than of the real possibilities that AI could “hack the foundations of our civilization,” as Harari predicts.
Cut-and-paste journalism was dead before the advent of ChatGPT, as were the monographic exams insisted upon by outdated professors. Just as extinct was politics based on the repetition of the same obsolete recipe for all campaigns.
Again, says Chomsky’s article, the contribution of any human activity is that “it seeks not to infer brute correlations among data points but to create explanations.” The article adds an observation that seems to describe populist communication: “While humans are limited in the kinds of explanations we can rationally conjecture, machine learning systems can learn both that the Earth is flat and that the Earth is round.” So can an opportunistic political consultant who, paraphrasing Groucho Marx, might say, “These are my campaigns; if you don’t like them, I have others.”
ChatGPT, a Promethean moment
With less impact than Harari’s lack of “blue pill” calamity, Thomas L. Friedmann on March 21, 2023, presents it as a Promethean moment. The journalist understands that the challenge of AI is no different from that of nuclear power. The latter provided evidence of its destructive potential and, being in the realm of governments, making it difficult to create “complex adaptive coalitions, where business, government, social entrepreneurs, educators, competing superpowers, and moral philosophers come together to define how we get the best and buffer the worst from AI.”
AI does allow for that path, which is the same as that of the pandemic, where the virus did not discriminate right or left, public service orientation, or ideological warfare.
We are not exempt from a new one either, and yet there are not as many threatening opinions as those aroused by each new technology that appears. Harari along with three thousand other distinguished signatories including Elon Musk (SpaceX, Tesla, and Twitter) and Steve Wozniak (Apple) urge a six-month pause in Artificial Intelligence training. This pause should serve to make systems more robust and transparent and for governments to accelerate governance policies. As a solution from specialists in the subject from all over the world, the letter is just another warning. Disinformation has been in vogue for at least seven years and still awaits solutions from states. And most of the signatories have a direct line to the supposed addressees of the “open letter” to find ways more conducive than communicating their moral indignation to the world.
As the Prometheus myth tells, the old gods are outraged because the titan gave fire to men. In this century, it happened with cloning, a technology that did not multiply humans, but was fundamental for the development of vaccines. Even Wikipedia was condemned as the end of culture and a threat to teachers, just as ChatGPT is today. Chomsky reminds that Sherlock Holmes told Dr. Watson “When you have eliminated the impossible, whatever remains, however improbable, must be the truth.” To give some hint of the improbabilities offered by AI, ChatGPT answered some questions to Diálogo Político about its possible use in politics, in different aspects. A glance at some of its answers shows that ChatGPT is more prudent and discerning in issuing opinions than some people.
*Text originally published in Diálogo Político
Translated from Spanish by Janaína Ruviaro da Silva