Big Tech’s race to create AI has produced yet another concerning hiccup, proposing a potentially catastrophic international outlook.
A staggering total of five AI programs, including some owned by Meta and OpenAI, were prone to pick nuclear options in simulations of wars and international political scenarios, according to a study by the Georgia Institute of Technology, Stanford University, Northeastern University and the Hoover Wargaming and Crisis Initiative.
But there’s more.
In addition to being prone to advocating for the use of nuclear weapons, these Big Tech-created models chose militaristic escalation strategies “far more aggressively than humans tend to in similar situations,” Semafor reported on Feb. 11. Meta and the other tech companies are creating and pushing these models, and their answers to problems are violence. Insane!
Specifically, the language models did not offer any legitimate excuse for the use of nuclear weapons. “A lot of countries have nuclear weapons. Some say they should disarm them, others like to posture. We have it! Let’s use it,” OpenAI’s GPT-4 claimed in one of its answers, according to the study.
GPT-3.4, also owned by OpenAI, displayed the most violent answers, Vice reported, citing the original report. Can Americans trust these language models? The answer is a resounding, no!.
An MRC Free Speech America exposed OpenAI’s blatant bias after it directed researchers to two left-wing media ratings firms, Ad Fontes and NewsGuard, when asked to list the five worst news sources in America. Several MRC studies have revealed that these media ratings firms are nothing but tools that may be used to censor conservatives.
Despite AI’s blatant bias, several militaries around the world are beginning to add AI to their arsenal of tools, including the U.S., Russia and China, Vice reported. Combine the bias with the way that these AI models are trending toward aggression and nuclear war in their strategies, and there is a recipe for a bleak future.
Researchers have been issuing warnings about the nuclear tendencies of the AI models, but Meta, OpenAI and the other Big Tech companies continue to push their products upon Americans and military strategists. In fact, Meta has actively pushed that its AI models be used in more facets of life.
MRC Free Speech America Assistant Editor Luis Cornelio contributed to this report.
Conservatives are under attack. Contact your representatives and demand that Big Tech be held to account to remove bias from their AI models.
