Artificial intelligence enthusiasts are betting that the technology can help solve humanity's biggest problems, from war to global warming, but that may actually be an unrealistic ambition at this point.
“It's not about asking AI, 'This is a tricky problem. What would you do?'” And AI is like, 'We need to completely restructure this part of the economy.' '' said Michael Littman, a professor of computer science at Brown University.
Littman had just attended South by Southwest (SXSW), an arts and technology festival in Austin, Texas, where he spoke on one of many panels about the potential benefits of AI.
“It's a pipe dream. It's a bit of science fiction. What people are mainly doing is trying to get AI to do certain problems that they're already solving, but they just want to make it more efficient.”
“It's not like you can just press this button and solve everything,” he says.
Promising titles (“How to Make AGI Beneficial and Avoid the Robot Apocalypse”) and the ever-present panels of tech giants draw large crowds, but are often aimed at more pragmatic purposes, such as promoting products. is listed.
At a conference titled “Inside the AI Revolution: How AI is enabling the world to do more,” Microsoft executive Simi Olabisi extolled the benefits of the technology in the company's cloud service, Azure.
With Azure's AI language capabilities in your call center, you can say, “Maybe when your customer called, they were angry, but when they got off the phone, they were really grateful.'' Azure AI Language “We can truly capture that emotion and tell businesses how their customers are feeling,” she explained.
～“Smarter than humans”～
The concept of artificial intelligence, with algorithms that can automate tasks and analyze large amounts of data, has been around for decades.
But last year, it took on a whole new dimension with the success of ChatGPT, a generative AI interface launched by OpenAI, a now-iconic AI startup primarily funded by Microsoft.
According to CEO Sam Altman, OpenAI claims it wants to build artificial “general” intelligence, or AGI, that is “generally smarter than humans” and “enhancing humanity.”
That spirit was very much on display at SXSW, where the talk was not about “if” but “when” AGI would become a reality.
Ben GoertzelScientists leading the SingularityNET Foundation and AGI Society have predicted the emergence of general AI by 2029.
“Even if we have a machine that can think like a smart human, it will take several years at most before we have a machine that can think a thousand or even a million times better than a smart human.'' “Because this AI can modify its own source code,” Goertzel said.
Wearing a leopard-print faux-fur cowboy hat, he said these “super AIs” were built into robots that were “like us” and had “compassion and empathy” to get along with humans. He advocated the development of AGI.
David Hanson, founder of Hanson Robotics and designer of Desdemona, a humanoid robot powered by generative AI, brainstormed the positives and negatives of superpowered AI.
“Positive disruption… by AI could help solve global sustainability problems, but people will probably only create absolutely effective financial trading algorithms,” he said. Ta.
Hanson is concerned about disruption caused by AI, but points out that humans have already done a “great job” in playing “existential roulette” with nuclear weapons and causing “the fastest mass extinction event in human history.” did.
But, he says, “there are seeds of wisdom in AI that can blossom and grow into new forms of wisdom that help us improve.”
– “We're not there yet” –
Advocates of AI initially said it should accelerate the design of new, more sustainable drugs and materials.
Even if “we're not there yet… In a dream world, AI will handle the complexity and randomness of the real world and… make things we never thought possible possible at all.” You can discover new materials.' It was possible,'' said Roxanne Talley, an investor at Piva Capital.
Today, AI has already proven its value, for example in tornado and forest fire warning systems.
However, in the event of a pandemic, it is still necessary to evacuate the population or get the population to consent to vaccination, said Raed Ghani of Carnegie Mellon University in a study titled “Can AI solve extreme weather pandemics?'' This was emphasized in a panel discussion.
“This problem is of our own making. Inequality was not caused by AI, it was caused by humans, and I think AI can help a little. But if humans use AI, only if we decide that we want to deal with it,” Ghani said.
juj/arp/tjj
