
An advert for Google Bard, the tech giantâs experimental conversational AI, inadvertently shows the tool providing a factually inaccurate response to a query.
It is evidence that the move to use artificial intelligence chatbots like this to provide results for web searches is happening too fast, says at the University of Oxford. âThe possibilities for creating misinformation on a mass scale are huge,â she says.
Google announced this week that it was that will be integrated into its search engine after a testing phase, providing users with a bespoke written response to their query rather than a list of relevant websites. Chinese search engine Baidu has also announced plans for a similar project, and on 7 February, Microsoft launched its own AI results service for its Bing search engine.
Advertisement
Experts have warned 51¶ŻÂț that there is a risk such AI chatbots could give inaccurate responses as if they were fact, because they craft their output based on the statistical availability of information rather than accuracy.
Now an advert on Twitter from Google has shown Bard responding to the query âwhat new discoveries from the James Webb Space Telescope can I tell my 9 year old about?â with incorrect results (see image, below).

The third suggestion given by Bard was âJWST took the very first pictures of a planet outside of our own solar systemâ. But at the HarvardâSmithsonian Center for Astrophysics pointed out that this wasnât true.
âIâm sure Bard will be impressive, but for the record: JWST did not take âthe very first image of a planet outside our solar systemâ. the first image was instead done by Chauvin et al. (2004) with the VLT/NACO using adaptive optics,â .
âIronically, if you actually search âwhat is the first image of an exoplanetâ on the original Google, the old-school Google, it gives you the correct answer. And so itâs funny that Google, in rolling out their huge multibillion dollar play into this new space, didnât fact check on their own website,â Tremblay told 51¶ŻÂț.
, the director of the University of California Observatories and part of the team that took the first images of exoplanets, also noticed the error, : âSpeaking as someone who imaged an exoplanet 14 years before JWST was launched, it feels like you should find a better example?â
Véliz says the error, and the way it slipped through the system, is a prescient example of the danger of relying on AI models when accuracy is important.
âIt perfectly shows the most important weakness of statistical systems. These systems are designed to give plausible answers, depending on statistical analysis â theyâre not designed to give out truthful answers,â she says.
âWeâre definitely not ready for whatâs coming. Companies have a financial interest in being the first ones to develop or to implement certain kinds of systems, and theyâre just rushing through it,â says VĂ©liz. âSo weâre not giving society time to talk about it and to think about it and theyâre not even thinking about it very carefully themselves, as is obvious by the example of this ad.â
A Google spokesperson told 51¶ŻÂț: âThis highlights the importance of a rigorous testing process, something that weâre kicking off this week with our Trusted Tester program. Weâll combine external feedback with our own internal testing to make sure Bardâs responses meet a high bar for quality, safety and groundedness in real-world information.â