Google subsidiary DeepMind has said it may soon launch a rival to ChatGPT, and its chatbot promises to be a safer type of AI assistant.
DeepMind has been at the forefront of AI research for the past decade and was acquired by Google 9 years ago. However, with ChatGPT in the headlines of late, DeepMind CEO Demis Hassabis told Time (opens in a new tab) that he plans to launch his own chatbot, called Sparrow, for a "private beta" in XNUMX.
Sparrow appeared on the planet last year as a completion test on a research paper (opens in new tab) that described him as a "helpful dialogue agent and one who reduces the danger of dangerous and inappropriate answers."
Despite some concerns about the potential risks of chatbots, which DeepMind says include "inaccurate or fabricated information," it looks like Sparrow could be ready to fly in beta soon. Given DeepMind's close relationship with Google, it could certainly become the search giant's answer to ChatGPT.
The slight delay in the release of Sparrow is due, according to Demis Hassabis, to DeepMind's concern to ensure that it has essential features that ChatGPT lacks, including citation of specific sources. As Hassabis told Time, "you have to be careful on that front."
From DeepMind's research work, it also appears that Sparrow will initially be more limited and conservative than ChatGPT. The latter went viral with his uncanny ability to assist everyone from coders to lounge versers, but also caused alarm with his ability to make discriminatory comments and his abilities to write malware.
DeepMind referred to the binding rules of behavior on which Sparrow relied, such as his willingness to refuse to answer questions in "contexts where it is appropriate to compromise with humans." In early tests, Sparrow reportedly provided a plausible answer and, more importantly, patently supported it "XNUMX% of the time when asked an objective question."
But its true capabilities will become clearer when this public beta releases later this year. Clearly we'll be holding the popcorn for the first ever AI chatbot discussion between Google-affiliated Sparrow and increasingly Microsoft-loving ChatGPT.
Analysis: AI chatbots are still in kindergarten
Anyone who has used ChatGPT will know that it can mimic intelligence on multiple subjects quite well. But while it's clearly a fun ride, AI chatbots also require ethical intelligence and the ability to refer sources, and that's where DeepMind claims its "dialogue broker" Sparrow is strongest. .
Taking it to the next level will require tons of outside input, so a public beta for Sparrow is imminent. DeepMind asserts that developing better rules for its AI assistant "will require both input from specialists in many subjects (including policymakers, social scientists, and moral specialists) as well as participatory input from a wide range of users and harmed sets."
Sam Altman, CEO of OpenAI (which created ChatGPT), also discussed the pitfalls of opening up AI chatbots without causing collateral damage. On Twitter (opens in a new tab), he agreed that "there are going to be significant issues with the use of OpenAI technology over time; we're going to do everything we can but won't successfully anticipate each and every one of them." drawbacks".
In other words, the developers of ChatGPT and DeepMind's Sparrow are like parents with curious little ones, which brings fun and risk in equal measure, especially when your kindergarten teacher is indeed all over the internet.
ChatGPT is already going crazy and is heading towards a monetized future with ChatGPT Professional, an imminent paid tier. But DeepMind's Sparrow looks like the sweeter character AI chatbots need as they move towards next-gen models, like the rumored ChatGPT-four.