AI is still a fairly unpredictable piece of tech, to the point that even the folks at Open AI don’t fully understand how to use it. This isn’t a particularly great look for the company, as they are supposed to be one of the leading faces of this futuristic technology.
In last week’s International Telecommunication Union AI for Good Global Summit, OpenAI CEO Sam Altman admitted that he still doesn’t completely understand AI (via Observer). Shockingly, Altman said "We certainly have not solved interpretability," before trying to claim that the technology is safe and attempting to compare human brains to it.
“We don’t understand what’s happening in your brain at a neuron-by-neuron level, and yet we know you can follow some rules and can ask you to explain why you think something,” said Altman.
Most people don’t understand how AI works, though some have come to know aspects of the technology. AI has to be trained so it only knows as much as the person telling it what to do. However, AI can interpret things differently based on its research, often stealing information and not giving users what they want.
Despite numerous claims by companies, it’s clear that this piece of technology isn’t ready to make any major decisions anytime soon. Various corporations have tried to tell fans that AI is the future, but when a conglomerate that’s supposed to specialize on this technology can’t predict it, no one will.
Not helping with all of this is the various controversies that come with using AI. There’s already plenty of art theft going around with the technology, one simply has to go online on Facebook or Twitter to see it. Identity theft is also a fear many logically have with AI, with ChatGPT getting sued by Scarlett Johansson for supposedly stealing her voice.
OpenAI will continue to try and make AI work, but it’s clear that they aren’t there yet. Sam Altman came back to the company not too long ago and things aren’t much better.