Will AI save humanity? Festival in the USA remembers its limits – 03/15/2024 – Tech

Fans of artificial intelligence (AI) are betting on this technology to help humanity solve problems such as global warming. But the reality is that this requires, above all, human efforts.

“Imagine we could ask AI ‘I have a difficult problem, what would you do in my place?’ and she replied ‘you must restructure a sector of the economy and then…’ No. It’s an illusion”, says Michael Littman, professor of computer science at Brown University, in the United States.

For him, artificial intelligence can help protect the environment, improving the efficiency of production systems, which will reduce energy consumption. “But it’s not just pushing a button. Humans will have a lot of work,” says he, who is participating in the SXSW arts and technology festival in Austin, Texas.

At the event, Microsoft executive Simi Olabisi promoted the company’s cloud services.

“The Azure Language tool for call centers allows you to capture the feelings of customers – who may, for example, be irritated at the beginning and very grateful at the end of the service – and report this to the company”, she explained at the conference “At the heart of the revolution of AI: How AI is Powering the World.”

Super AI

The notion of artificial intelligence, with its algorithms capable of automating tasks and analyzing large amounts of data, has existed for several decades.

However, it took on a new dimension last year with the success of ChatGPT, the generative AI interface launched by OpenAI, a company primarily funded by Microsoft.

The technology, which produces text, images and other content in plain language, raises concerns about employment or fraud, but it also generates a lot of excitement.

OpenAI (in addition to Google, Meta, Amazon) wants to create a “general” AI (AGI) that works based on programs “smarter than humans in general”, mainly to “help elevate humanity “, according to its CEO, Sam Altman.

For Ben Goertzel, scientist who leads the SingularityNET Foundation and the AGI Society, he estimates that the tool will be launched in 2029.

“When we have a machine that thinks as well as an intelligent human being, it will only take a few years until the machine thinks a thousand times better, or a million times better, than a human being, because it will be able to modify its own source code,” he said during the conference “How to make general AI beneficial to avoid the robotic apocalypse?”.

Goertzel advocates the development of general AI endowed with “compassion and empathy” and integrated with humanoid robots “that look like us”, to ensure good understanding between future “super AI” and humanity.

‘Seeds of Wisdom’

For Hanson Robotics founder David Hanson, AI could “help solve environmental problems, although people will probably use it to create financial trading algorithms”, he joked.

His company designed the humanoid robot Desdemona, equipped with generative AI, which was also present at the event.

However, Hanson fears the potentially devastating consequences of a widespread AI race across countries, but notes that humans have not waited for this technology to appear “to play existential roulette with nuclear weapons” or “cause a mass extinction.”

Perhaps general AI “will contain seeds of wisdom that will blossom and help us improve,” he said.

Initially, this tool should allow us to accelerate the design of new medicines or more sustainable materials.

“If we dream a little, we can understand the complexity of the physical world and use it to […] discover completely new materials that allow us to do things that were previously impossible,” Roxanne Tully, from the Piva Capital investment fund, explained to AFP.

AI is already giving good results in forecasting and warning systems in the event of tornadoes or forest fires, but it is still necessary to remove affected populations or for humans to agree to be vaccinated in the event of a pandemic, highlighted Rayid Ghani, from Carnegie Mellon University. .

“We create the problems, not AI. Technology can help us… a little (…) only if humans decide to use it to solve these problems”, he concluded.