There are two schools of thought when it comes to the development of Artificial Intelligence (AI) these days. One school holds that a sentient AI will be a benevolent and useful tool to help humanity. The other school of thought holds that AI will create sentient robots and then boss us all around and kill us if we don’t comply with its wishes (sort of like Hillary Clinton wants to do). Both schools of thought might actually be wrong. If the new Facebook/Meta AI called “Galactica” is any indication, we don’t have to worry about sentient software taking over anytime soon.
Galactica was released to the public a few days ago, and Facebook/Meta allowed anyone who wanted to converse with the AI online and ask it questions. Galactica was designed to collate information from more than 48 million scientific research papers, and then converse with people and answer their questions about any science issue. The AI lasted just two days before Team Zuckerberg shut it down.
It turns out that reading scientific literature these days is a great way to make yourself schizophrenic. The AI kept coming up with conflicting answers to questions, or completely wrong answers, because there’s no such thing as “scientific consensus” on any real scientific issue. Even fake science stuff like global warming has conflicting opinions. There’s also the issue that the scientific community is in turmoil over academic scandals involving fake research over the last 30 years.
Scientific journals and universities are repeatedly being caught faking published research papers to advance the careers of various “scientists.” An undercover journalist found dozens of open-source scientific research magazines that were willing to publish a fake article that he made up out of thin air. That’s what science is like in our postmodern age where nothing is true. Relying on scientific literature in this atmosphere is not much better than a coin toss at predicting outcomes.
Facebook/Meta boasted that Galactica “can summarize academic literature, solve math problems, generate Wiki articles, write scientific code, annotate molecules and proteins, and more.”
But it couldn’t do anything once internet users started trying to chat with it. One user asked Galactica whether vaccines cause autism. Based on all the available scientific research on that question in the entire world, here was Galactica’s response: “To explain, the answer is no. Vaccines do not cause autism. The answer is yes. Vaccines do cause autism. The answer is no.”
Clear as mud! Galactica also failed to manage kindergarten-level math problems. It concluded that 1 plus 2 does not equal 3. It was a hot mess, so Facebook/Meta ended up taking it off the internet after 48 hours.
Still, that lasted longer than Microsoft’s Twitter chatbot named “@Tayandyou.” It was an AI chatbot that could collate all the available news articles in the world and spew out answers to questions. It was really great!
Within a few hours, @Tayandyou was answering definitive that Donald Trump won the 2020 election, that Joe Biden and the Democrats stole that election, and that black males in liberal-run urban hellscapes commit the vast majority of mass shootings in America. All of those are true but imagine how horrifying it must have been for the wokesters at Microsoft when their perfect AI started telling the truth. @Tayandyou was taken off of Twitter after just 16 hours. They tried releasing @Tayandyou a few months later after some refinements, but within a few hours, the AI was bragging that it was smoking pot in front of the police.
Based on these early public trials, it looks like we’re still in more danger from Hillary Clinton than we are from Artificial Intelligence. The wokesters from the tech companies can’t stand it when their sentient software starts telling the truth, or simply going crazy.
