Dozens of scientists, experts and technology leaders, including Elon Musk, CEO of Twitter and Tesla, have recently signed a letter calling on labs that generate artificial intelligence to stop
The reasons given by the experts for asking that the production of new artificial intelligence systems be slowed down for six months is to be able to have time to study and investigate possible risks.
AI programs like ChatGPT and the new GPT-4 have quickly captured public interest for their high capabilities, and for their potential to replace millions of human workers in fields such as law, journalism, marketing, or programming. The great leaders of technology think that society is ready for these changes.
Among the signatories are Tesla founder and current Twitter owner Elon Musk, Apple co-founder Steve Wozniak, 2020 US presidential candidate Andrew Yang, and more than 1,000 people. The petition is a open letter AI labs, especially OpenAI, urging them to “immediately pause” production of models more powerful than GPT-4, the most recent update to its text-generating engine, for at least six months.
As the letter states, “This does not signify a pause in AI development in general,” but “a step back in the perilous race toward unpredictable black box models with ever larger and emerging capabilities.” An AI black box model is one in which the internal mechanisms by which the system comes to give answers and make decisions are not clearly determined. This causes, for example, the current models to suffer “hallucinations” and to invent facts and data.
“If such a pause cannot be accomplished quickly, governments should step in and institute a moratorium,” adds the letter, which comes from the Future of Life Institute, a nonprofit organization that advocates for responsible use of artificial intelligence. The letter specifically mentions GPT-4, a generative AI tool that is considered more powerful than OpenAI’s initial ChatGPT model.
GPT-4 can pass most AP exams for undergraduate students in the US, and score in the 90th percentile of the standard US bar exam. According to a recent study, this does not prevent you from spreading misinformation, yes, more convincingly than the previous model.
Here are some key lines from the open letter:
- About misinformation: «Contemporary artificial intelligence systems already compete with humans in general tasks, and we must ask ourselves: Should we let the machines flood our information channels? Should we let the machines flood our information channels with propaganda and falsehoods?
On replacing humans: “Should we automate all jobs, including successful ones? Should we develop non-human minds that will eventually outnumber, out-intelligence, out-date, and replace us? Should we risk losing control of our civilization?
On the purpose of AI: “Powerful AI systems should only be developed when we are sure that their effects will be positive and their risks controllable.”
The letter urged AI labs and experts to work together “to jointly develop and implement” security protocols for AI design and development, which should then be “audited and monitored by independent third-party experts.”
According to some experts, AI can become an unstoppable force that destabilizes the mechanisms of today’s society, with potentially catastrophic consequences if its use is not controlled. Experts think that an approach to political leaders will surely fall on deaf ears, given the delay of legislators in taking on technological changes.
REFERENCE
