
Meta CEO Mark Zuckerberg has claimed that his company’s latest artificial intelligence systems have shown early signs of self improvement.
In a policy paper on the future of AI, the tech founder wrote that one of the key benchmarks for superintelligence – a general AI that is smarter than humans – is within sight.
“Over the last few months we have begun to see glimpses of our AI systems improving themselves,” he wrote. “The improvement is slow for now, but undeniable.”
One of the main fears surrounding superintelligence, frequently expressed by academics and industry leaders, is that a self-improving system could lead to out-of-control and rogue AI.
Mr Zuckerberg did not address these particular concerns, but did warn that advanced artificial intelligence comes with serious risks.
“We believe the benefits of superintelligence should be shared with the world as broadly as possible. That said, superintelligence will raise novel safety concerns,” Mr Zuckerberg wrote.
“We'll need to be rigorous about mitigating these risks and careful about what we choose to open source. Still, we believe that building a free society requires that we aim to empower people as much as possible.”
He claimed that his firm’s approach to superintelligence differs from its rivals, as it is focussing on empowering individuals rather than using it to automate “all valuable work”.
With billions of people around the world using Meta products like Facebook, Instagram and WhatsApp, the billionaire said that his company is in the best position to integrate AI into people’s daily lives.
If successful, the result will be people spending “less time in productivity software, and more time creating and connecting”, he said.
His memo came just ahead of the company’s quarterly earnings, which saw better-than-expected financial results following massive investment in AI.
Earnings of $7.14 per share on $47.52 billion in revenue surpassed Wall Street expectations of $5.92 in earnings per share on revenue of $44.8bn.
Meta’s share price rose by double digits, as investor appeared hopeful that the tech giant’s multi-billion dollar gamble on AI development will pay off.
“In some ways, this [superintelligence] race is historically akin to the races for the PC, web browser, search engine, and smartphone,” Mike Proulx, a research director at Forrester, told The Independent.
“But the big difference is that this race is pacing so much faster because AI – the very thing that Meta and others are racing towards – helps to accelerate itself.”