Human-like intelligence by 2030: AI could 'destroy mankind', Google warns

New York: A new research paper by Google DeepMind has predicted that the rapid progress in artificial intelligence could ‘permanently destroy humanity’ by 2030.

The study pointed out that the human-level artificial intelligence (AI), popularly called Artificial General Intelligence (AGI), could pose existential risk to humanity ‘Given the massive potential impact of AGI, we expect that it too could pose potential risk of severe harm’, NDTV reported. 

While AI is task-specific, AGI aims to take it further forward allowing the technology to possess intelligence capable of applying in ‘wide range of tasks’ much like human intelligence.

AGI will emerge with the ability to ‘understand, learn, and apply knowledge in diverse domains, much like a human being’.

The report comes amid growing concerns over AI technology invading several areas of social life including taking over jobs.

The paper, co-authored by DeepMind co-founder Shane Legg, said that ‘the question of whether a given harm is severe isn't a matter for Google DeepMind to decide’ adding that ‘it is the purview of society, guided by its collective risk tolerance and conceptualisation of harm’.

Rather than specify how AGI is going to cause extinction of humankind, the study zeros in on the preventive measures AI companies including Google should take to tackle the AGI’s threat.

The study categorizes potential risks of advanced AI into four groups namely misuse, misalignment, mistakes and structural risks alongside highlighting DeepMind’s risk mitigation strategy of focusing on prevention of misuse of AI.

Demis Hassabis, CEO of DeepMind, called for setting up a UN-like organization to monitor AGI’s development while predicting that AGI that are smart or smarter than humans would emerge in the next five or ten years.

‘I would advocate for a kind of CERN for AGI, and by that, I mean a kind of international research focused high-end collaboration on the frontiers of AGI development to try and make that as safe as possible,’ Hassabis was quoted as saying.