Google subsidiary DeepMind has said it may soon launch a ChatGPT rival – and its chatbot promises to be a safer type of AI assistant.
DeepMind has been a pioneer in AI research for the past decade and was acquired by Google nine years ago. However, with ChatGPT stealing recent headlines, DeepMind CEO Demis Hassabis told Time (opens in a new tab) that it plans to release its own chatbot, called Sparrow, for a “private beta” in 2023.
Sparrow was introduced to the world last year as a proof of concept in a research paper (opens in a new tab) who describes it as a “useful dialogue agent and one that reduces the risk of dangerous and inappropriate responses”.
Despite some concerns about the potential dangers of chatbots, which DeepMind says include “inaccurate or made-up information,” it looks like Sparrow could be ready to take flight soon in beta 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 is, according to Demis Hassabis, due to DeepMind’s concern to ensure that it has important features that ChatGPT lacks, including citing specific sources. As Hassabis told Time, “you have to be careful on that front.”
From DeepMind’s research paper, it also appears that Sparrow will initially be more constrained and conservative than ChatGPT. The latter went viral with his impressive ability to help everyone from coders to parlor poets, but he also alarmed with his ability to make discriminating comments and his skills in writing malware.
DeepMind referred to the binding behavioral rules that Sparrow relied on, as well as his willingness to refuse to answer questions in “contexts where it is appropriate to defer to humans.” In early tests, Sparrow apparently provided a plausible answer and, crucially, backed it up with evidence “78% of the time when asked a factual question.”
But its true capabilities will become clearer when this public beta launches later this year. We’ll definitely be grabbing the popcorn for the first AI chatbot debate between the Google-affiliated Sparrow and the increasingly Microsoft-loving ChatGPT.
Analysis: AI chatbots are still in kindergarten
Anyone who has used ChatGPT will know that they are able to do a pretty good imitation of intelligence on a number of topics. But while it’s definitely a fun ride, AI chatbots also need moral intelligence and an ability to cite sources – and that’s where DeepMind says its “dialogue agent” Sparrow is strongest. .
Taking it to the next level will require tons of external input, which is why a public beta of Sparrow is imminent. DeepMind claims that developing better rules for its AI assistant “will require both input from experts on many topics (including policymakers, social scientists and ethicists) and participatory input from a wide range of affected users and groups”.
Sam Altman, CEO of OpenAI (which created ChatGPT), also spoke about the difficulties of opening AI chatbots without causing collateral damage. At Twitter (opens in a new tab) he admitted, “there will be significant problems with the use of OpenAI technology over time; we will do our best but we will not successfully anticipate all problems”.
In other words, the developers of ChatGPT and DeepMind’s Sparrow are like parents with curious toddlers, which brings fun and danger in equal measure, especially when their kindergarten teacher is effectively all internet.
ChatGPT is already going wild and heading towards a monetized future with ChatGPT Professional, an imminent paid tier. But DeepMind’s Sparrow seems like the sweetest character AI chatbots need as they move towards next-gen models, like the rumored ChatGPT-4.