The MIT author of a significant letter warns that unbridled AI development, controlled by a few companies, jeopardizes society’s future
A few tech firms endanger humanity’s future with unchecked AI progress. The scientist leading a pivotal letter urges them to cease the “race to the bottom.” Max Tegmark, a physics professor and AI expert at MIT, warns of a concerning trend: “We’re observing an alarming descent that requires intervention.” Tegmark orchestrated an open letter, signed by numerous tech leaders like Elon Musk and Apple’s Steve Wozniak, advocating a six-month pause on large-scale AI experiments.
Tegmark emphasized, “We’re observing an alarming descent that must be halted. We urgently require AI safety standards to turn this into a race for excellence. While AI offers numerous incredible benefits, the reckless and unrestrained advancement of increasingly potent systems without oversight endangers our economy, society, and lives. Regulation is imperative for secure innovation to prevent a few AI giants from compromising our collective future.”
In a recent policy document, 23 AI experts, including two prominent pioneers of the field, argued that governments should have the authority to pause the development of exceptionally powerful AI models.
Gillian Hadfield, a paper co-author and the head of the Schwartz Reisman Institute for Technology and Society at the University of Toronto, stated that AI models currently under development are projected to be significantly more potent than the ones currently in use within the next 18 months.
The paper, authored by Geoffrey Hinton and Yoshua Bengio, both recipients of the ACM Turing Award, often referred to as the “Nobel prize for computing,” argues that highly advanced AI models should necessitate government licensing and, if required, be subject to development suspension.
“For exceptionally capable future models, such as those with the potential to surpass human control, governments should be prepared to authorize their development, temporarily halt their progress if concerning capabilities arise, enforce access controls, and demand robust information security measures capable of thwarting state-level hackers, until sufficient safeguards are in place.”
The unbridled advancement of artificial general intelligence, denoting a system capable of performing a wide array of tasks at or above human intelligence levels, remains a significant concern for proponents of stricter regulation.
Recently, Amazon declared an investment of up to $4 billion in Anthropic, a startup founded by former OpenAI executives, aiming to leverage proprietary data hosted on Amazon’s cloud services. The precise extent of Amazon’s ownership in Anthropic is undisclosed, but it is expected to be a minority stake in the company, previously valued at $4 billion. Notably, Anthropic is currently facing a lawsuit from music labels for alleged copyright infringement related to the use of song lyrics.
On the contrary, Meta, the parent company of Facebook, has shifted its focus from the metaverse to AI, allocating a $33 billion budget for the current year to bolster the expansion of AI capabilities. Their primary focus is on advancing their open-source Llama set of large language models.
Mark Zuckerberg, CEO of Facebook, mentioned during an earnings call in April that Meta envisions introducing AI agents to billions of individuals in practical and meaningful ways. The company plans to concentrate its AI investments in two key areas. Firstly, they aim to enhance the extensive recommendations and ranking infrastructure supporting their primary products, ranging from news feeds to short video reels, their advertising system, and integrity mechanisms, which have been under development for many years. Secondly, they plan to focus on new generative foundation models that enable entirely novel categories of products and experiences.
Meta’s strong performance in ad revenue, reaching $28 billion for its “family of apps” segment (including Facebook, Instagram, Messenger, and WhatsApp), is attributed to the increased utilization of AI recommendations. Consequently, the value of Meta’s shares has more than doubled in 2023.
The report suggests that advancements in generative artificial intelligence have the potential to trigger profound transformations within the global economy.
Additionally, there is a significant surge in investments in new AI companies, surpassing other technology investment sectors, reaching nearly $18 billion in the third quarter. This represents a 27% global increase compared to the corresponding period in the previous year.