OpenAI’s ChatGPT is all the rage right now, a chatbot that can provide answers in a human-like conversational tone to complex questions. That’s why companies like Microsoft and Google scrambled to take advantage of the ChatGPT momentum.
Microsoft seems to be in the lead, having included the actual ChatGPT in its Bing app. Google announced the ChatGPT-like Bard would be coming soon, but its chatbot already made an error in an ad for the upcoming service.
That mistake cost the company $100 billion in market value, as the shares tanked 7% after word got out. But Bard’s mistake is actually a great development all ChatGPT users have to pay attention to and learn from. Yes, generative AI is amazing, and ChatGPT-like products are the future. But we’re still in the imperfect present, and these bots aren’t ready to replace traditional online search products.
Google’s first public reaction to ChatGPT was that it couldn’t include a similar tool in Search, as the chatbot can make mistakes. In turn, this would put its reputation in danger. But Google began scrambling to shift resources towards its own ChatGPT-like rivals immediately.
A few days ago, we heard that Bard was already in testing internally, just as Google teased AI announcements for some of its core apps, including Maps, Search, and Translate. Then Google CEO Sundar Pichai unveiled Bard officially on Monday. Microsoft released ChatGPT-enabled Bing to the world.
Then, Google on Wednesday held its AI event in Paris, just as a report came out detailing Bard’s first error. The ChatGPT-like product had to answer this question: “What new discoveries from the James Webb Space Telescope can I tell my 9 year old about?”
One of the answers the AI chatbot provided was that the telescope “took the very first pictures of a planet outside of our own solar system.” But that’s not factually correct, something astrophysicist Grant Tremblay pointed out on Twitter. He also showed that Google Search delivered the correct answer to the question.
That Google’s Chatbot rival made such a big mistake is perhaps one of the best things about the dawn of this generative AI frenzy. We’ll see these bots everywhere in the coming months and years. And they’ll get to a point where they will offer correct, reliable answers. But we’re not quite there yet. Google isn’t there yet.
It’s very likely that Bard is more sophisticated than OpenAI’s ChatGPT. A report last week hinted at that, saying that Bard had access to more recent information than ChatGPT when formulating its answers in internal tests.
These AI tools are still imperfect, however. They can make mistakes, like the one in the Bard ad that briefly tanked Google’s shares.
Let’s not forget that ChatGPT made mistakes from the start and continues to make them. If you weren’t aware of the fact, the Google Bard mishap should drive that point home. The chatbots are in their infancy, and there’s a lot of work left to ensure they’ll become the reliable helpers these tech companies want them to be.
For the moment, you shouldn’t trust them completely, not when you know their answers can be wrong. Or that ChatGPT-like bots can be manipulated with questions to produce incorrect content.
The best way to realize that the chatbots are still very early software solutions isn’t ChatGPT making a mistake, something OpenAI warned from the start. It’s Google’s Bard making mistakes right from the get-go. Precisely because you don’t expect mistakes from Google Search. While we’re at it, also expect Bing’s ChatGPT to make errors along the way.
So why are Microsoft and Google hurrying to deploy unfinished AI products? Just to make a clear statement. Both companies were there at the start of this new tech age. They have ChatGPT-like products, and they’re willing to perfect them so they can become truly amazing.
Until then, expect mistakes, including the kind that go viral and tank stocks. It’s all par for the course. Just remember to fact-check the answers using traditional search results.
