Big Tech is scrambling to release products that can compete with OpenAI’s ChatGPT, the sensational AI chatbot that is also the fastest growing app ever. Google’s entry, Bard, to be released “in the coming weeks”(opens in a new tab)According to a Google blog post, but it is already thwarting ChatGPT’s impersonation by generating false information.
Google’s blog entry about Bard features an animated graphic designed to demonstrate the Bard user experience, and long story short, it contains an AI falsely claiming that the James Webb Space Telescope took the first picture of an exoplanet. Google also tweeted an animation with this statement.
Tweet may have been deleted
(opens in a new tab)
(opens in a new tab)
Webb took over last September Its first photo of an exoplanet(opens in a new tab)But it wasn’t the first picture of an exoplanet — a milestone that happened in 2004(opens in a new tab),
It’s unclear what happened, but one eyebrow-raising aspect of Bard’s claim about James Webb is how recent it is. The information in a language model is not pulled from a list of facts and read in, but is extracted by very sophisticated systems to complete sentences. Sentences about the very recent past may be more flawed than usual for an AI because the information in them simply hasn’t been written that often. This is probably one of the reasons why ChatGPT’s model won’t tell you much about the year 2021(opens in a new tab),
This snafu speaks to the longstanding problem of all generative AI products disregarding truth value, which is potentially a good reason for consumers to continue using old-fashioned search engines. Microsoft is in the process of integrating a ChatGPT-like answer engine into its search engine, Bing, but as Microsoft notes(opens in a new tab)“Bing sometimes misrepresents the information it receives, and you may see responses that seem reassuring but are incomplete, inaccurate, or inappropriate.”
Thanks, but for now I think I’ll just use google. er, the Old Google.
