Google’s DeepMind says it may soon release a competitor to ChatGPT — and its chatbot promises to be a safer kind of AI assistant.
DeepMind has been a leader in AI research over the past decade and was acquired by Google nine years ago. However, with ChatGPT stealing recent headlines, DeepMind CEO Demis Hassabis told Time magazine: (Opens in a new tab) It is considering releasing its chatbot, named Sparrow, for a “private beta” sometime in 2023.
Sparrow was introduced to the world last year as a proof of concept in a research paper (Opens in a new tab) which she described as “a useful dialogue agent that reduces the risk of unsafe and inappropriate answers”.
Despite some skepticism about the potential dangers of chatbots, which DeepMind says contain “inaccurate or invented information,” it looks like Sparrow may soon be ready to fly in experimental form. Given DeepMind’s close relationship with Google, it could effectively become the search giant’s answer to ChatGPT.
The slight delay in the launch of Sparrow, according to Demis Hassabis, is due to DeepMind’s care to ensure that it has important features that ChatGPT lacks – in particular, citing specific sources. As Hassabis told Time magazine, “It is right to be careful on this front.”
From DeepMind’s paper, it also seems that Sparrow will initially be more restrictive and conservative than ChatGPT. The latter has gone viral with its remarkable ability to help everyone from programmers to arm-poets, but has also caused concern with its ability to discriminate comments and malware-writing skills.
DeepMind has spoken of the behavior restraining rules Sparrow built on, along with its willingness to refuse to answer questions in “contexts in which it is appropriate to defer to humans.” In early tests, Sparrow gave a seemingly plausible answer, crucially backing it up with evidence “78% of the time when asked a factual question.”
But its true capabilities will become clearer when the public beta launches later this year. We’re sure to be grabbing popcorn for the first AI chatbot debate between Google’s Sparrow and Microsoft’s increasingly lovable ChatGPT.
Analysis: AI chatbots are still in kindergarten
Anyone who has used ChatGPT will know that it can do very good imitations of intelligence on a number of topics. But while this is certainly a fun ride, AI chatbots also need moral intelligence and the ability to cite sources — and this is where DeepMind says its Sparrow ‘dialogue agent’ is strongest.
Taking this to the next level is going to need a lot of outside input, which is why Sparrow’s public beta is so imminent. DeepMind says that developing better rules for its AI assistant “will require input from experts on many subjects (including policy makers, social scientists, and ethicists) and participatory input from a variety of affected users and groups.”
Sam Altman, CEO of OpenAI (which created ChatGPT), similarly spoke of the difficulties in unlocking AI chatbots without causing collateral damage. on me Twitter (Opens in a new tab) He admitted, “There will be significant problems with using OpenAI technology over time; we’ll do our best but we won’t anticipate every problem successfully.”
In other words, the developers of both ChatGPT and DeepMind’s Sparrow are like parents with curious kids, bringing fun and danger in equal measure—especially when the kindergarten teacher is all the Internet.
ChatGPT is already running wild and heading into a monetized future with ChatGPT Professional, an imminent paid class. But it looks like DeepMind’s Sparrow might be the milder character that intelligent chatbots need as they race toward next-generation models, like the rumored ChatGPT-4.