AI Chatbot Provides Quack Advice To Make Cheese Stick To Pizza

Google’s AI chatbot got itself into more embarrassing blunders — this time “teaching” people to use non-toxic glue to make cheese stick to pizza.

The BBC reported its experimental “AI Overviews” tool has told some users searching for how to make cheese stick to pizza better that they could use “non-toxic glue”. The search engine’s AI-generated responses have also said geologists recommend humans eat one rock per day.

In a bid to save reputational damage (whatever is left), Google spokesperson reportedly claimed these blunders were “isolated examples”. Some data were created by trolls on Reddit, a platform notorious for its unpopular content, or pieces from The Onion. 

With the vast amount of data inputted into the software known as training, one could not help but wonder if the world is training the AI chatbot to be stupid instead. Despite the bad press, Google insisted the feature was generally working well.

“The examples we’ve seen are generally very uncommon queries, and aren’t representative of most people’s experiences. The vast majority of AI overviews provide high quality information, with links to dig deeper on the web.” it said in a statement.

In one shocking example reported by BBC, a reporter Googled if they could use gasoline to cook spaghetti faster was told “no… but you can use gasoline to make a spicy spaghetti dish” and given a recipe. Google said it had taken action where “policy violations” were identified and was using them to refine its systems.

Despite these incidents which questioned the credibility of AI chatbot’s answers, a lawmaker actually deemed ChatGPT an expert in law. An Arizona state representative behind a new law that regulates deepfakes in elections used an artificial intelligence chatbot, ChatGPT, to write part of the law – specifically, the part that defines what a deepfake is.

Republican Alexander Kolodin’s bill, which passed unanimously in both chambers and was signed by the Democratic governor recently, will allow candidates in Arizona or residents to ask a judge to declare whether a supposed deepfake is real or not, giving candidates a way to debunk AI-generated misinformation.

Kolodin said he used the chatbot ChatGPT to help define what “digital impersonation” is for the bill in part because it was a fun way to demonstrate the technology. He provided a screenshot of ChatGPT’s response to the question of what a deepfake is, which is similar to language that is included in the bill’s definition.

Back to making pizza, would you trust a chatbot’s recommendation “non-toxic glue” to make cheese stick to the pizza dough?

What You Missed:

Arizona State Lawmaker Used ChatGPT To Write Part Of Law On Deepfakes
ASML Secret Sauce For Semiconductor Success Amid Challenges In The Angstrom Era
AI Has Crept Its Way Into Aerial Combat
Tesla Profits Decline By More Than 50% In Q1 2024
Tesla Reduces Full Self-Driving Software Price To US$8,000
Researchers Turn Metal Waste Into Catalyst For Hydrogen
Is AI Is Looking More Like A Band Aid Now?
Rever Automotive Builds And Tests BYD Buses In Thailand
Tesla Cuts More Than 10% Of Its Workforce
Tesla Settling Lawsuit Erodes Faith In Autonomous Driving

WANT MORE INSIDER NEWS? SUBSCRIBE TO OUR DIGITAL MAGAZINE NOW!

CONNECT WITH US:  LinkedIn, Facebook, Twitter

Letter to the Editor
Do you have an opinion about this story? Do you have some thoughts you’d like to share with our readers? APMEN News would love to hear from you!

Email your letter to the Editorial Team at [email protected]