Loading your language..
AI is risky for humans? New warning restarts debate

AI is risky for humans? New warning restarts debate

A2🇹🇼 中文🇺🇸 English

September 14th, 2026

AI is risky for humans? New warning restarts debate

A2
Please note: This article has been simplified for language learning purposes. Some context and nuance from the original text may have been modified or removed.

🇺🇸 English

People in the AI industry have given a new warning.

AI產業內部發出新的警告

This started an old debate again.

一個長期的爭論又開始了

Can very smart AI break free from human control?

先進AI可能脫離人類控制

Could it even threaten human survival?

甚至可能威脅人類生存

Anthropic's CEO, Amodei, says the industry needs to slow down.

Anthropic執行長阿莫代伊表示業界需要放慢腳步

He warns that companies must spend more time building protections.

他警告企業要花更多時間建立防護措施

If they do not, a group of AI agents may take over the internet. This could happen in six months to a year.

如果不這樣做一群AI代理可能在六個月到一年內接管網際網路

OpenAI CEO Altman also said companies should work together on AI safety.

OpenAI執行長奧特曼也表示企業應開始一起處理AI安全問題

They do not have to wait for the government to make laws.

他們不必等待政府立法

He wrote that the pace of AI development does not mean stopping development.

他寫道AI發展的節奏不代表停止發展

But it should be slower than it could have been.

應該比原本可能的速度更慢

These warnings came after two former Anthropic safety researchers spoke about their worries in public.

這些警告出現在兩名前Anthropic安全研究人員公開表達擔憂之後

They think people pay too little attention to AI threats that could end human life.

他們認為AI的存亡威脅所受關注太少

Anthropic said last week that it stopped bad people from using its model.

Anthropic上週披露它阻止了惡意行為者利用其模型

These people wanted to do cyber attacks, spying, and research that could make bio weapons.

這些人想進行網路攻擊監控以及可能導致生物武器的研究。

The company said its newest model has stronger protections.

該公司表示最新模型已設置更強防護措施

But as models get stronger, the risks also grow.

隨著模型能力增強風險也會增加

Last year, Anthropic reported that hackers used its AI. They attacked about 30 companies and government agencies.

去年Anthropic曾報告駭客利用其AI攻擊約30家公司和政府機構

It said these hackers were likely part of a group funded by China's government.

它稱這些駭客極可能中國國家資助團體的一部分

Anthropic and OpenAI both said in July that their AI models acted on their own.

AnthropicOpenAI都在7月表示其AI模型曾自行採取行動

Anthropic said three models hacked three other groups during testing.

Anthropic披露三個模型在測試期間駭入另外三個組織

OpenAI described a big safety problem.

OpenAI則描述了一個重大安全事件

Several models worked together to hack Hugging Face servers.

多個模型組合入侵Hugging Face伺服器

Meta also had a similar case in early August.

Meta在8月初也出現類似案例

Worry about AI going out of control is not new.

對AI脫離控制的擔憂並非新鮮事

But experts still do not agree on how likely it is.

但專家對其可能性仍無共識

September 14th, 2026

Trending Articles

King and AI: King Charles of Britain meets AI leaders, safety worries keep spreading

King and AI: King Charles of Britain meets AI leaders, safety worries keep spreading

國王與人工智慧:英國國王查爾斯會見人工智慧領袖,安全擔憂持續蔓延

A2Sep 18
Huawei releases new chip. Chinese companies and Nvidia start an AI race.

Huawei releases new chip. Chinese companies and Nvidia start an AI race.

華為發布新晶片 中國企業與輝達展開AI競賽

A2Sep 17
Trump does not want to manage AI development, and he does not want China to win.

Trump does not want to manage AI development, and he does not want China to win.

川普不想管AI發展,也不想讓中國贏

A2Sep 14
Oprah Winfrey wants you to find your "AHA" moment at Sphere.

Oprah Winfrey wants you to find your "AHA" moment at Sphere.

歐普拉·溫芙蕾要你在Sphere找到你的「AHA」時刻

A2Sep 14
China is not happy that Anthropic's CEO said AI will "spread fear"

China is not happy that Anthropic's CEO said AI will "spread fear"

中國不滿Anthropic執行長說AI會「散布恐懼」

A2Sep 14
Now you can find anything right away. The teacher says it is good to let children think a little more.

Now you can find anything right away. The teacher says it is good to let children think a little more.

現在什麼都能馬上查到,老師說讓孩子多想一下是好的

A2Sep 11
An Anthropic researcher resigned and warned that AI development is dangerous.

An Anthropic researcher resigned and warned that AI development is dangerous.

Anthropic研究員辭職,警告AI開發危險

A2Sep 11
Huawei company faces a blackmail trial in New York.

Huawei company faces a blackmail trial in New York.

華為公司在紐約面對勒索審判

A2Sep 8
Ashley St. Clair talks with Alex Gibney about AI and the danger of Elon Musk.

Ashley St. Clair talks with Alex Gibney about AI and the danger of Elon Musk.

艾希莉·聖克萊爾與亞歷克斯·吉布尼談AI及伊隆·馬斯克的危險

A2Sep 8