惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

N
Netflix TechBlog - Medium
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
爱范儿
爱范儿
博客园_首页
雷峰网
雷峰网
Hugging Face - Blog
Hugging Face - Blog
V
Visual Studio Blog
The Cloudflare Blog
罗磊的独立博客
美团技术团队
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
大猫的无限游戏
大猫的无限游戏
阮一峰的网络日志
阮一峰的网络日志
V
V2EX
博客园 - 叶小钗
The GitHub Blog
The GitHub Blog
Last Week in AI
Last Week in AI
J
Java Code Geeks
MyScale Blog
MyScale Blog
G
Google Developers Blog
U
Unit 42
Y
Y Combinator Blog
P
Proofpoint News Feed
Vercel News
Vercel News

Artificial Intelligence News -- ScienceDaily

The shape behind the Einstein problem just revealed strange new physics AI uncovers hidden Ozempic side effects across 400,000 Reddit posts Tiny nanolaser could cut computer energy use in half Scientists just made quantum computer operations 1,000 times faster A “quantum bath” puts quantum entanglement on autopilot IBM quantum computer solves classically intractable problem in 15 minutes NASA just used satellites and debris to navigate without GPS Tiny robots powered by light can hunt down and collect bacteria Scientists turn DNA into a memory device that uses 100x less power Scientists tracked kids for 8 years — the screen time result was unexpected World’s first superconducting quantum heat engine could help unlock massive quantum computers MIT’s new lidar chip could give self-driving cars a wider view New programmable photonic chip can control how fast light moves Alan Turing's biggest AI assumption may have been wrong Scientists discovered the brain doesn't make decisions the way we thought Quantum mechanics once baffled scientists. Now it's changing the world Millions of exploding stars could soon reveal dark energy's secrets SpaceX wants to build AI data centers in space. Will it work? Brain-inspired chip runs near absolute zero and could transform quantum computing Scientists are seriously asking if bees and ChatGPT are conscious Forget electrons, this breakthrough uses light-matter particles to power AI NASA’s new AI space chip could let spacecraft think for themselves New quantum algorithm solves “impossible” materials problem in seconds Your “um” and pauses could reveal early dementia risk AI lets chemists design molecules by simply describing them This AI knew the answers but didn’t understand the questions AI swarms could hijack democracy without anyone noticing Think AI "knows" what it’s doing? Scientists say think again Artificial neurons successfully communicate with living brain cells Quantum AI just got shockingly good at predicting chaos
A classic brain test exposed AI's biggest weakness
2026-06-10 · via Artificial Intelligence News -- ScienceDaily

Artificial intelligence systems can write essays, answer questions, and solve complex problems. But new research suggests they may struggle with something humans do every day: staying focused on the task at hand when distractions get in the way.

Researchers led by Suketu Patel put several leading AI models through a well-known psychology experiment called the Stroop task. The results revealed a significant difference between how AI systems process information and how the human brain manages attention.

What Is the Stroop Task?

The Stroop task is a classic psychological test that has been used for decades to study attention, concentration, and self-control.

In the test, color words such as "red," "blue," or "green" are displayed in colored ink. Sometimes the word and the ink color match. For example, the word "red" might appear in red ink. Other times they conflict, such as the word "red" printed in blue ink.

Participants are asked to name the color of the ink rather than read the word itself.

That sounds simple, but it creates a challenge because reading words is an automatic habit for most people. The brain must suppress the urge to read the word and instead focus on identifying the ink color.

Psychologists often use the task to measure what is known as executive control, a set of mental processes that helps people regulate attention, resist distractions, and stay focused on goals.

Testing AI Attention

The researchers wanted to see whether modern large language models (LLMs) handle this challenge in the same way humans do.

LLMs are the AI systems behind tools such as ChatGPT, Claude, and Gemini. They are trained on enormous amounts of text and learn patterns in language, allowing them to generate responses that often appear remarkably human.

When given short lists containing five color words, the AI systems generally performed well, even when the words and colors did not match.

However, the picture changed dramatically as the lists became longer.

GPT-4o achieved 91% accuracy when working with five words. At ten words, its accuracy fell to 57%. When the list expanded to forty words, accuracy dropped to just 15%.

Claude 3.5 Sonnet maintained stable performance through lists of twenty words but then experienced a sharp decline, falling to 24% accuracy with forty-word lists.

The researchers observed similar patterns in GPT-5, Claude Opus 4.1, and Gemini 2.5.

When AI Loses Focus

The challenge became even more difficult when matching and mismatched color words appeared together in the same list.

Under those conditions, performance deteriorated further. Accuracy for the mismatched items dropped to nearly zero in some cases.

According to the researchers, the AI models had trouble maintaining the instruction to identify ink colors. Instead, they increasingly defaulted to reading the words themselves.

In other words, the systems appeared unable to consistently suppress the response they had been most heavily trained to produce.

This finding is particularly interesting because humans face a similar conflict. People are generally much better at reading words than naming ink colors. Yet despite this bias, most individuals can maintain high accuracy and stable performance even when confronted with long lists of conflicting words and colors.

Human Attention vs. Machine Attention

The study highlights an important distinction between human and artificial intelligence.

Although modern AI systems can produce impressive language and reasoning capabilities, their underlying mechanisms differ from the attention processes found in biological brains.

Humans can often sustain focus on a specific goal while filtering out competing information. The results suggest that current AI models may struggle with this type of cognitive control when tasks become increasingly demanding.

The researchers argue that the performance collapse seen in these experiments points to fundamental limitations in today's large language models. While AI can sometimes mimic human behavior, its ability to maintain attention appears to operate very differently from the way people do.

The findings offer a reminder that even the most advanced AI systems still have weaknesses, particularly when tasks require them to resist distractions and stay focused over extended sequences of information.