For a brief moment in 2023, the biggest existential threat to humankind seemed to come not from anthropogenic climate change but another human-driven specter: artificial intelligence. Ushering in this new dystopian flavor was OpenAI’s launch of ChatGPT-3, a large language model, or LLM, capable of generating long sequences of predictive text based on human prompts.
In the following weeks, as people across the world consumed billions of watts of energy submitting prompts like “Rewrite the Star Wars prequels where Jar Jar Binks actually is the secret Sith Lord,” the public discourse became congested with a slurry of technological prognostications, philosophical thought experiments, and amateurish science-fiction plots about sentient machines.
Major media outlets published op-eds like “Can We Stop Runaway A. I.?” and “What Have Humans Just Unleashed?” Governments across the West scrambled to form oversight committees, and every tech-savvy person adopted a new vernacular almost overnight, exchanging terms like “machine learning” and “data science” for “AI.”
While OpenAI’s launch incited an LLM arms race among tech titans like Google, Amazon, and Meta, some prominent technologists, like Elon Musk and Steve Wozniak, signed an open letter warning about the ominous future of an unchecked artificial intelligence, urging all AI labs to halt their experiments until our regulatory apparatuses (and our ethics) could catch up. On the pixelated pages of the New York Times and Substack, public intellectuals openly grappled with moral quandaries posed by the specter of an omnipotent artificial intelligence.
Although AI’s most zealous boosters and fearful doomers both likely overstated the capabilities of LLMs and the velocity of…
La suite est à lire sur: jacobin.com
Auteur: Jennifer Lenow
