The Tesla boss has been warning of the consequences of artificial intelligence for some time.
(Photo: IMAGO/ZUMA Wire)
Dusseldorf It’s a “Who’s Who” of thought leaders from the tech scene: Apple co-founder Steve Wozniak, Skype co-founder Jaan Tallinn, historian and best-selling author Yuval Noah Harari and Stuart Russell, professor of computer science at the University of Berkeley, are all there. As the “head of SpaceX, Tesla and Twitter”, Elon Musk is also one of the first to sign an open letter on the subject of artificial intelligence (AI), which is quite a challenge.
In the letter released on Wednesday, more than 1,000 experts, entrepreneurs and scientists demand that “all AI labs immediately stop training AI systems that are more powerful than GPT-4 for six months”. GPT-4 is the current version of the AI developed by the start-up OpenAI. This is behind the chatbot ChatGPT, which has been causing a sensation for months.
The letter was initiated by the “Future of Life Institute”, a US foundation founded by Musk, Tallinn and others in 2014, which, according to its statutes, is “dedicated to reducing existential risks, especially from advanced AI”.
The signatories cite a “profound risk for society and humanity” as the reason. AI has gotten so good that it can do tasks like humans. This harbors the danger that “fulfilling jobs will be automated away”, that “we will be flooded with propaganda and untruths” and that AI is not only smarter, but “outnumbers” and could replace humanity.
“Should we risk losing control of our civilization?” the letter reads.
ChatGPT: Sharp criticism of OpenAI
OpenAI is criticized without naming the start-up: With the publication of its AI via ChatGPT, the company has started a “dangerous race” for a technology that “no one – not even its inventors – understands, can predict or reliably control”. .
GPT stands for “Generative Pre-Trained Transformer”. This is an artificial intelligence that can write high-quality texts and generate images or sounds. It is based on the so-called Transformer model, which uses neural networks and data training to predict and order the following words in an incomplete sentence or individual pixels in an unfinished image according to their probability.
More about artificial intelligence
OpenAI released the fourth version of GPT a few days ago. Microsoft has a stake in the company and, according to media reports, will invest around ten billion dollars in the coming years. As a result, other leading companies such as Google or Baidu were forced to launch their own chatbots and AI systems.
Elon Musk is co-founder of OpenAI
Musk’s concerns about AI are nothing new. Years ago he warned about the technology and therefore founded OpenAI together with Sam Altman and Peter Thiel in 2015, at that time still as a foundation. This should ensure that AI is developed openly and transparently and is equally accessible to everyone. However, Musk departed in 2018 when Altman formed a for-profit subsidiary of the same name and brought Microsoft on board.
“We need some kind of regulator or someone to oversee AI development,” Musk said at Tesla Investor Day a few weeks ago. The letter also calls for “security protocols for advanced AI design and development that are rigorously reviewed and overseen by independent and outside experts.”
Critics accuse Future of Life of fueling the hype surrounding this technology with the open letter. The call only serves to stoke fears, said Johanna Björklund, AI researcher and professor at the University of Umea. “There is no reason to pull the handbrake.” Instead, the transparency requirements for developers should be tightened.
With agency material
More: ChatGPT maker OpenAI presents improved version GPT-4
First publication: 03/29/2023, 5:39 p.m.