Mark Zuckerberg broadcasts plans for AGI, raises concerns

January 20, 2024: Mark Zuckerberg has drawn criticism over his plan to develop a high-tech company advanced artificial intelligence (AI) system. The goal of this AI is to succeed in and even exceed human intelligence levels.

Zuckerberg also hinted at the potential of making this technology open source, meaning anyone could use it.

As CEO of Meta, Zuckerberg announced the corporate’s ambitions to construct a man-made general intelligence (AGI) system. AGI refers to a theoretical AI able to performing a big selection of tasks with intelligence equal to or exceeding that of humans.

The idea of ​​AGI has sparked fear amongst experts and leaders all over the world, who fear such a system could escape human control and change into a threat.

Zuckerberg explained that the subsequent wave of technology services requires the creation of complete general intelligence.

Meta’s plan is to potentially share its AGI technology with external developers and the general public. He emphasized that this technology ought to be made available as widely as possible but in a responsible manner.

Dame Wendy Hall, a professor on the University of Southampton and an advisory member of UN AI, described the thought of ​​open source AGI as “very scary indeed”. She criticized Zuckerberg for even considering this step.

According to Hall, Releasing such powerful AI without proper regulation might be harmful. However, she expects that it’s going to take a few years for AGI to be fully developed, leaving time for the mandatory regulations to be established.

Meta, amongst other tech corporations, has agreed to let governments review the safety of AI tools before and after they’re released. This commitment was made at a world AI security summit within the UK.

Dr. Andrew Rogoyski from the University of Surrey identified that the choice to open source an AGI system mustn’t be left to 1 technology company alone. He stressed the necessity for international consensus on such crucial issues.

Zuckerberg recently mentioned in an interview that he would lean towards open sourcing – so long as it stays protected and responsible. However, Meta’s previous decision to open source their Llama 2 AI model was met with criticism. Some experts likened it to providing a blueprint for an atomic bomb.

Other major AI players equivalent to OpenAI and Google DeepMind are also working on AGI.

Sam Altman, CEO of OpenAI, said on the World Economic Forum that AI advances rely on energy breakthroughs equivalent to nuclear fusion.

While Zuckerberg didn’t provide a timeline for AGI’s development, he did mention Meta’s extensive infrastructure and ongoing work on a successor to Llama 2, underscoring the corporate’s serious commitment to this ambitious AI project.