The one simple question that proves AI won’t save us: ChatGPT fails VERY basic test – despite scientists banking on technology for early cancer detection

The one simple question that proves AI won't save us: ChatGPT fails VERY basic test - despite scientists banking on technology for early cancer detection

As an ardent follower of advancements in technology and a survivor of a battle against cancer myself, I must express my disappointment with the current state of Artificial Intelligence, particularly after hearing about the debacle involving ChatGPT and their inability to correctly identify the number of “R”s in the word “strawberry.”


Scientists are relying on Artificial Intelligence as a valuable tool in their fight against cancer.

Yet, an Australian breakfast radio team has challenged the ChatGPT AI system with a simple query; surprisingly, the answer was mishandled in a rather dramatic fashion.

During their radio show on Thursday morning (The Kyle & Jackie O Show), the hosts, Kyle Sandilands and Jackie O Henderson, chose to challenge the capabilities of the AI model, ChatGPT, live on-air. Their experience with it was so unexpected that they were taken aback by the results.

The question posed to the program developed by Open AI, named ‘strawberry’, couldn’t accurately provide the number of ‘R’ letters it contains.

In response to the query, the app mistakenly stated that there are only two “R” sounds in the given word instead of the actual count of three.

Despite repeated corrections, the software continues to make the same mistake as it learns from user interactions, suggesting it has not fully grasped the correct information yet.

Kyle, 53, quipped, ‘We’re not going to be curing cancer any time soon with this thing are we?’

During the live broadcast, the hosts showed a viral video of the app committing a spelling error and requested their program director, Bruno Bouchet, to test it on the spot.

‘How many Rs are in strawberry?’ asked Bruno, before the chatbot replied, ‘there are two Rs.’

The one simple question that proves AI won't save us: ChatGPT fails VERY basic test - despite scientists banking on technology for early cancer detection

The studio then gasped in shock as others began laughing.

Afterward, Bruno endeavored to evaluate the AI’s ability to learn by commenting, “It seems to me that there are three.”

‘I see where you’re coming from,’ said the robotic voice, before making a wildly inaccurate claim. 

In the term “strawberry,” it may seem as though there are numerous instances of the letter R, but in truth, it contains just two such sounds: one right after the ‘T’ and another at the very end.

Following unsuccessful efforts to teach the AI, the radio personalities expressed dissatisfaction regarding the excessive attention given to such educational programs.

‘You stupid piece of s**t,’ Kyle said, before adding ‘I thought this thing would cure cancer.’ 

The test comes just months after Scottish scientists revealed they have developed an Artificial Intelligence system that can interpret cancer samples and predict how a tumour might progress.

The one simple question that proves AI won't save us: ChatGPT fails VERY basic test - despite scientists banking on technology for early cancer detection

The team believes this finding might empower physicians with an artificial intelligence’s additional perspective, potentially accelerating and improving the precision of diagnostic procedures.

A group of AI experts and cancer researchers, with the University of Glasgow at the helm, have created a novel method, referred to as histomorphological phenotype learning (HPL), in their collaborative efforts.

Researchers acquired and recorded high-quality images of numerous tissue specimens from lung cancer patients, also tracking the course of their respective cancer developments.

The one simple question that proves AI won't save us: ChatGPT fails VERY basic test - despite scientists banking on technology for early cancer detection

Afterward, they created an algorithm designed to scrutinize the images and identify recurring patterns strictly using the visual information contained within each slide.

The algorithm divided the slide images into countless minute fragments, with each fragment symbolizing a tiny portion of human tissue.

As the team incorporated additional slides of cancer patients into the system, it became proficient at accurately identifying those patients’ distinctive characteristics.

After recognizing patterns within the given samples, researchers proceeded to apply the algorithm for examining connections between these samples and the health results of the patients, specifically focusing on post-surgery survival times for those who had undergone cancer operations.

Another recent study claimed AI was better at detecting prostate cancer than doctors. 

Read More

2024-08-29 04:29

Previous post Watch the top 3 fastest knockouts in each UFC division
Next post Star Trek: Deep Space Nine writers didn’t get Worf but thankfully Michael Dorn did