Can you get secondhand embarrassment for an entire company? Because that’s what watching Google‘s Live From Paris event felt like.
On the heels of Microsoft’s exciting news about the new AI-powered Bing, Google’s Bard AI technology was reported to be inaccurate and faced event blunders. It’s an uncharacteristic occurrence for the tech giant and underscored the “code red” Google declared internally following the meteoric popularity of ChatGPT.
To be fair, the main purpose of Google’s event was to announce new Google Maps and Google Search features and updates that have been in the works for a while. Immersive View and Glanceable Directions are some of the standouts that would’ve been the big highlights in different circumstances. But ChatGPT’s integration into the new and improved Bing forced Google to say something about Bard, which overshadowed the rest of the event and drew attention to Google’s shortcomings compared to OpenAI’s advanced conversational AI technology.
Join us as we recap the mess that unfolded.
Bard didn’t do its fact-checking
As the event was happening, Reuters published a report revealing Bard’s inaccuracy. In a promo video on Google’s blog showing off Bard’s capabilities, the AI falsely said the James Webb Space Telescope took the first ever picture of an exoplanet. It was Webb’s first picture of an exoplanet, but the first picture taken occurred back in 2004.
“Information in a language model is not pulled from lists of facts and recited, but spat out by very sophisticated systems for completing sentences,” Mashable’s Mike Pearl reported. “Sentences about the very recent past could plausibly be more error-riddled than usual for an AI because the information in them simply hasn’t been written as many times.”
In other words, this is a common issue for all current AI language models, not just Bard. But for this report to come out as the event was happening was not a good look for Google, which prides itself on its AI.
A demo phone was…misplaced
To be clear, this is not about ridiculing the poor presenter who just experienced everyone’s public-speaking nightmare. But it speaks to the overall haphazard tone of the event. During the announcement about a new Lens feature, the demo phone was misplaced and the presenter wasn’t able to show the demo, leaving her literally empty-handed and forced to wrap up the news with a promise to show the feature in the Q and A that followed.
The presenter showed admirable composure, but how was this allowed to happen? This was one of the features that Google had been planning to announce, so it’s not like the presenter was told to hop on stage at the last minute. Was the Google event so chaotic that a phone was lost? Why wasn’t there a backup phone?
Anyone who has put on an event knows that mistakes like this do sometimes happen, but it was an unfortunate blunder during an already bizarre event that had me wondering if Google needs a wellness check.
The Bard non-announcement
When Google announced it was hosting an event this week about search and AI, the internet eagerly began speculating about how it would respond to the threat of ChatGPT. Then, the company officially launched Bard on Monday which ramped up the anticipation for more details about the conversational AI technology.
Instead, we were given a glorified repeat of what was already announced. The company did share a few details about how Bard will power Google search results and will be open to testers next month. But, it felt like a last-minute addition that was shoehorned into the pre-planned event. Attendees expecting a worthy rebuttal to Microsoft’s Bing announcement were left underwhelmed.
The livestream is no longer live
After all that, Google seemed to say, “let’s pretend this never happened,” and immediately made the livestream recording private after the event.
Given the opportunity to flex its AI muscles and alleviate concerns about being spooked by ChatGPT, Google made the situation worse by appearing panicked and reactive. Until now I didn’t think ChatGPT was a truly credible threat to Google’s AI technology. Maybe OpenAI got there first, but Google would surely come out with something even more powerful to top it.
That may still happen in the coming weeks, but this event shows that Google’s worst enemy isn’t ChatGPT, but itself.
