The seemingly sudden advancement of artificial intelligence (AI) has sparked concern across the globe. Will everyone's jobs be replaced? Are degrees now useless? Will a bunch of random computer coding one day turn its back on humanity? While there are some serious consequences of developing a machine that can talk, write, think, and read like a human - there are also some positives to the newly-emerged tech, with scientists already working alongside AI to help improve cancer detection rates.

But more work needs to be done to refine these powerful machines, which once suggested adding non-toxic glue to a pizza base to help the cheese stick, recently went into a meltdown after being asked a rather simple question...

Swathes of social media users discovered that most major AI models struggled to correctly answer how many 'Rs' are in the word 'Strawberry'. The system would come back stating 'two' instead of 'three' and even doubled down when users tried correcting it, adding: "I apologise for the confusion, but I must respectfully disagree. The word 'strawberry' does contain two 'Rs'."

It appears the blunder has now been changed, but not before hundreds of baffled users started joking about AI's power to rule over us. "I think we're safe," one person quipped on X (formerly known as Twitter). Another agreed, commenting: "AI has street smarts, not book smarts."

A third added: "So they don't actually listen and give a half-assed response? Maybe AI is closer to us than we thought," while a fourth penned: "While even a kindergartener can get this right, the most advanced AI models struggle with this simple counting task."

Want to avoid the doom and gloom? Get the latest positive news sent straight to your inbox with our Bright Stuff Newsletter

Taking to TikTok, user Alberta Tech explains the glitch is due to how AI's language model works - showing how most models will assign every single world you type to them to a token ID (a numeric digit). "Then, the prediction that the AI model makes is based on these numbers," the software engineer states. "It's not based on the actual text input.

"So the AI model never gets to actually see the word strawberry. So when you ask how many R's are in the word strawberry - it is making what us computer scientists call a random freaking guess."

Do you think AI is the future? Have your say in the comments section below

Sign up to FREE email alerts with news to brighten your day Never miss any of the fun stuff. Get the biggest stories and wackiest takes from the Daily Star, including our special WTF Wednesday email

We have more newsletters

Never miss any of the fun stuff. Get the biggest stories and wackiest takes from the Daily Star, including our special WTF Wednesday email

We have more newsletters

Looks like a strawberry a day will keep the robots away...

Widespread fears artificial intelligence (AI) and its ubiquitousness will rule over the world have been quelled, after most major models completely glitched at a basic question a toddler could easily answer. The newly-emerged tech, which is being trained to think, speak, write, and read like a human being, has divided the nation - with many insisting the machines will steal our jobs, let our kids cheat on their homework, and become a deadly part of modern warfare.

Others have welcomed the tech, praising AI for helping to improve cancer detections and even predict long-term risks of the disease. But throw a fruit at it, and AI isn't happy...

READ MORE: 'Killer robots' will enforce 'grim future' on humans if the world doesn't unite

Click here for the latest bizarre news from the Daily Star

A slew of social media users discovered that most major AI models incorrectly answered the question: "How many Rs are there in Strawberry?"

Instead of replying with 'three' as you would expect, the machine insisted there were only 'two'. When challenged, AI doubled down on its reply - stating: "I apologise for the confusion, but I must respectfully disagree. The word 'strawberry' does contain two 'Rs'."

Straight away, Brits started roasting the idea of AI taking over the world, with one person stating "I think we're safe." Another quipped: "AI has street smarts, not book smarts," While a third added: "Seems like it could be a pretty powerful deterrent for using AI to answer questions you don't know the answer to."

User Alberta Tech, a software engineer, took to TikTok to explain why the blunder - which appears to now be solved - occurred. She points to the language model, which assigns every single world you type in to a token ID (a numeric digit).

"Then, the prediction that the AI model makes is based on these numbers," she said. "It's not based on the actual text input. So the AI model never gets to actually see the word strawberry. So when you ask how many R's are in the word strawberry - it is making what us computer scientists call a random freaking guess."

For the latest breaking news and stories from across the globe from the Daily Star, sign up for our newsletters.