- Google Bard AI and Microsoft ChatGPT-powered Bing made mistakes in their demos due to hallucination, which occurs when AI makes up incorrect details.
- The AI hallucination problem may go unnoticed if they are not tasking with crawling the internet in real-time, which exposes them to fake information and other issues.
- Both companies are aware of the errors and are working to improve their systems.
Google Bard AI and Microsoft ChatGPT-powered Bing both made mistakes in their demos due to something called hallucination.
When an AI like this large language models hallucinate, they make up wrong details. It does not mean that they have a consciousness that makes them dream.
Why does this happen? The technology is built on prediction or an autocomplete system which analyze patterns and determine what words should come next in a sequence, rather than looking at a database of the facts and bring up answers from there.
The AI hallucination problem may go unnoticed if both Google and Microsoft are not tasking Bard and ChatGPT, respectively, to start scouring the internet as search engines and making up answers. With the task of crawling the internet in real-time, these systems are exposed to fake information and other problems.
–Tesla Autopilot Workers Write Elon Musk About Their Union Plans
–Nigeria’s Curacel Raises $3 Million In Seed Funding
–Elon Musk Tweets Are On Everyone’s Timeline But A Fix Is Coming.
According to a CNBC report, the Bing tool integrated with ChatGPT made several errors when it was asked to analyze earnings reports from clothing brands Gap and Lululemon.
The media outlet reported that the AI reported the gross margin of Gap wrongly by not removing Yeezy charges while making up the company’s operating margin totally. It made similar errors for Lululemon as well.
As for Google Bard AI, it wrongly claimed that the James Webb Space Telescope took the very first pictures of a planet outside of our own solar system. When in reality, the first such image was taken in 2004.
Both companies are aware of the errors that their AI systems made. Google, through its spokesperson, said it is starting a rigorous testing process to use external feedback to improve Bard. Meanwhile, Microsoft, eager to push out their offering, advices users to check the facts provided by its tool and share feedback as well.
For your daily dose of tech, lifestyle, and trending content, make sure to follow Plat4om on Twitter @Plat4omLive, on Instagram @Plat4om, on LinkedIn at Plat4om, and on Facebook at Plat4om. You can also email us at firstname.lastname@example.org and join our channel on Telegram at Plat4om. Finally, don’t forget to subscribe to OUR YOUTUBE CHANNEL.