We’re far from human-level A.I., early DeepMind investor says

We’re still a long way from human-level AI despite rapid advances in technology, according to an early investor in research lab DeepMind.

“On AI, OpenAI and ChatGPT stuff: It’s like saying we’re going to jump to the moon,” said Humayun Sheikh, a founding investor in AI startup DeepMind, which is now owned by Googleto CNBC in an interview.

Related investment news

“We’ve made a big leap, but we’re not even on the moon yet.”

Sheik, who owned about 1.3% of DeepMind stock in 2011, said large language models (LLMs) like those developed by Microsoft-backed OpenAI, while impressive, are a far cry from so-called general intelligence. artificial, or AGI.

Artificial general intelligence is often referred to as the holy grail of artificial intelligence. It is a virtual system capable of completing any task at the same level as a human.

“This is really how I compare AI to all the big language model companies that have come out,” said Humayun, who is now the co-founder of AI and blockchain startup Fetch.ai.

“It’s very limited. How to actually get them to do certain things is still in its infancy.”

“It’s too much: Google isn’t born yet, but Yahoo is,” he added.

His comments come as Alphabet, Google’s parent company, has merged DeepMind with Google Brain, part of the US-based internet giant’s research division.

Google is racing to compete with Microsoft and other technology companies in the field of artificial intelligence. Microsoft is making great strides with its investment in OpenAI and inclusion of the company’s LLM technology in the Bing browser and other products.

Earlier this week, Demis Hassabis, founder of DeepMind, said, he told the Wall Street Journal that some form of artificial general intelligence might be possible “in the next few years”.

Al-Sheikh said he had “a lot of respect” for the accounts and that the entrepreneur was “very knowledgeable about AI ethics.”

“One of the first things he always brought up was how do we control it? How do we put those boundaries around it and make sure AI doesn’t get out of control?” Sheikh said.

Google acquired DeepMind for $500 million in 2014 and is trying to boost its business by doubling down on artificial intelligence in a bid to fend off the threat to its core search unit from OpenAI. Google launched its chatbot replacement for ChatGPT, Bard, in March.

The huge potential of AI is seen in its ability to generate entirely new content from user prompts. People have used this technique to create everything from poems to whimsical pictures and movie clips, while kids use it to help with their homework.

Experts have raised concerns about the dangers of cutting-edge AI, however, and a group of technology pioneers, including Elon Musk and Steve Wozniak, have called for a six-month ban on developing AI more advanced than GPT-4, the latest version of OpenAI. The huge one. language processing software.

He watches: Could Chinese ChatGPT clones give it an edge over the US in an AI arms race?