“有图有真相”的时代要结束了吗?

近日，OpenAI发布AI生成图片新模型GPT-Image-2，引发热议。

注：以下图片均为AI生成。

今天你练听力了吗？

🤔️小作业：

1. What is the primary purpose of the article?

A. To promote OpenAI's new image model by highlighting its realistic capabilities.

B. To criticize OpenAI for blurring the line between real and fake images.

C. To analyze the legal implications of AI-generated images on copyright law.

D. To explain how ChatGPT Images 2.0 uses thinking capabilities to improve accuracy.

无注释原文：

OpenAI wants you to know how good its new image model is at faking real photos

From: Business Insider

OpenAI wants you to know how good its latest image generator is —and it has a few fake screenshots and pictures of pictures to prove it.

On Tuesday, OpenAI announced ChatGPT Images 2.0, a new image model that it says has "thinking capabilities," which includes the ability to crawl the internet and "double check" its own output.

Some of the examples are made to look like fake screenshots or of a person thumbing through a magazine, further blurring the lines between what's real and fake as AI-generated images and videos continue to proliferate on the internet.

In teasing the release, OpenAI shared an AI-generated picture on X that mimics a desktop screenshot of a Google Chrome tab on a Mac. The fake screenshot shows a tab opened to ChatGPT's website, talking about Tuesday's livestream of the new model.

"The team really cooked on this one," OpenAI CEO Sam Altman said during the livestream.

OpenAI said in a press release that Images 2.0 can also generate pictures from prompts in multiple languages, including Japanese, Korean, Chinese, Hindi, and Bengali.

In samples provided by OpenAI, the model created realistic advertisements, a photo of what appears to be a physical magazine, and multiple pages of manga.

A spokesperson for OpenAI said the company does not intend to copy or reproduce specific artwork and that the images are generated from patterns it has learned.

Outside mimicry, the model allowed for some creative twists on real-life celebrities and images, like an 1980s glam version of legendary investor Warren Buffett and a blueprint of a Business Insider editor's dinner.

The spokesperson added that OpenAI prevents copying styles of individual living artists, but it does allow recreating broader studio styles.

Mitch Stoltz, the IP litigation director for the Electronic Frontier Foundation, told Business Insider in an interview that the latest model alone doesn't raise copyright issues, and that the distinction OpenAI makes between images generated from specific artwork and those generated from patterns the model learned is important.

"If the output is substantially similar to something that the model was trained on or crawled, then there starts to be a copyright concern," Stoltz said. "But if it's similar on the level of an idea, or just something that occurs in the world, or even a style or a vibe — generally speaking, that's not enough, and that's not a copy under copyright law."

OpenAI is facing at least a dozen ongoing copyright suits from writers and news outlets, including The New York Times and George R.R. Martin.

"The copyright issues are the same as if you were simply using Photoshop, a darkroom, or a human artist to make these kinds of replicas," he said. "The issues for society, whether you call that ethical or political or whatever, are greater because it's easier, faster, and more available."

- ◆ -

注：完整题目见本文开头；中文文本为机器翻译，仅供参考

含注释全文：

OpenAI wants you to know how good its new image model is at faking real photos

From: Business Insider

OpenAI wants you to know how good its latest image generator is —and it has a few fake screenshots and pictures of pictures to prove it.

OpenAI想让你见识一下它最新图像生成模型的厉害——为此，它特意拿出了几张伪造的截图和“画中画”来证明。

screenshot /ˈskriːn.ʃɒt/ 表示“屏幕截图”，英文解释为“an image created by copying part or all of the display on a computer screen at a particular moment, for example in order to demonstrate the use of a piece of software”

On Tuesday, OpenAI announced ChatGPT Images 2.0, a new image model that it says has "thinking capabilities," which includes the ability to crawl the internet and "double check" its own output.

周二，OpenAI发布了ChatGPT Images 2.0，这是一个全新的图像模型，据称具备“思考能力”，能够浏览互联网并“复核”自己的输出结果。

crawl /krɔːl/ 1）作动词，表示“缓慢移动；艰难前行；爬行；匍匐前进”，英文解释为“to move slowly or with difficulty, especially with your body stretched out along the ground or on hands and knees”举个🌰：The child crawled across the floor. 那个孩子爬过了地板。

2）表示“爬取”或“抓取”。在计算机和互联网领域，crawl（通常称为web crawling，即网络爬虫）是指程序自动浏览互联网上的网页、读取内容并收集信息的过程。

3）作名词，crawl /krɔːl/ 表示“自由泳；爬泳”，英文解释为“a style of swimming in which you move your arms over your head and kick with straight legs”

double-check /ˌdʌb.əlˈtʃek/ 表示“复核，复查”，英文解释为“If you double-check something, you make certain it is correct or safe, usually by examining it again.”

Some of the examples are made to look like fake screenshots or of a person thumbing through a magazine, further blurring the lines between what's real and fake as AI-generated images and videos continue to proliferate on the internet.

其中一些示例被设计成看似伪造的截图，或是一个人翻阅杂志的画面。随着AI生成的图像和视频在互联网上不断泛滥，这进一步模糊了真实与虚假之间的界限。

thumb /θʌm/ 表示“匆匆翻阅，浏览”，英文解释为“to turn the pages of a book, magazine, or a document quickly and only read small parts of it”举个🌰："Have you read the report?" "Well, I thumbed through it quickly while I was eating breakfast." “你看过报告了吗？”“嗯，我在吃早饭时匆匆翻阅过。”

blur /blɜːr/ 作动词，表示“（使）变模糊”，英文解释为“to (make something or someone) become difficult to see clearly”或者表示“使…的区别不明显；使难以了解…的真相”，英文解释为“to make the difference between two things less clear, or to make it difficult to see the exact truth about something”举个🌰：This film blurs the line/distinction/boundary between reality and fantasy. 这部电影模糊了现实与幻想的界限。

作名词，表示“模糊形状；模糊的东西”，英文解释为“a shape that you cannot see clearly, often because it is moving too fast”举个🌰：If I don't wear my glasses, everything is just a blur. 如果不戴眼镜，我看任何东西都是模糊的。

proliferate /prəˈlɪf.ər.eɪt/ 表示“激增”，英文解释为“to increase a lot and suddenly in number”举个🌰：Small businesses have proliferated in the last ten years. 在过去的10年里，小型企业大量涌现。

In teasing the release, OpenAI shared an AI-generated picture on X that mimics a desktop screenshot of a Google Chrome tab on a Mac. The fake screenshot shows a tab opened to ChatGPT's website, talking about Tuesday's livestream of the new model.

在为发布做预告时，OpenAI在X平台上分享了一张AI生成的图片，模仿了Mac电脑上谷歌浏览器标签页的桌面截图。这张伪造的截图显示，一个标签页打开了ChatGPT的网站，内容正是关于周二新模型的直播。

tease此前在iPhone 12为啥没发布？一文中出现：Google teased two new 5G smartphones. tease常见含义是：1）表示“嘲笑；取笑”，英文解释为“To tease someone means to laugh at them or make jokes about them in order to embarrass, annoy, or upset them.”举个🌰：He teased me mercilessly about going Hollywood. 他无情地取笑我想去好莱坞。

2）表示“梳理”，英文解释为“If someone teases their hair, they separate the individual strands from each other, for example by combing it.”举个🌰：Her hair was teased until it stood out and around her face. 她的头发被梳得直垂在脸旁。

📍貌似两个意思都和文中的解释不通，Wiktionary中最后一条释义最为贴切：To show as forthcoming, in the manner of a teaser.

tease xxx表示的是，预告，预热xxx，可以是事情/活动/产品之类的。此前在How You Like That...文中，我们就提到另一个词，teaser，意思是“前导广告，悬念式广告（含蓄而引人好奇）”，英文解释为“an advertisement for a product that does not mention the name of the product or say much about it but is intended to make people interested and likely to pay attention to later advertisements”，比如各种MV影片的“预告”，“预热”片，就可以说是teaser。

mimic /ˈmɪm.ɪk/ 1）表示“模仿（人的言行举止）；（尤指）做滑稽模仿”，英文解释为“to copy the way sb speaks, moves, behaves, etc., especially in order to make other people laugh”举个🌰：She's always mimicking the teachers. 她总喜欢模仿老师的言谈举止。

2）表示“（外表或行为举止）像，似”，英文解释为“to look or behave like sth else”举个🌰：The robot was programmed to mimic a series of human movements. 机器人可按程序设计模仿人的各种动作。

"The team really cooked on this one," OpenAI CEO Sam Altman said during the livestream.

“团队在这件事上真是大显身手，”OpenAI首席执行官萨姆·奥尔特曼在直播中说道。

在这个语境下，“cook” 是俚语用法，意思是“做得非常出色”、“状态极佳”、“拿出了惊艳的表现”。这种用法多出现在科技圈、游戏圈或年轻人之间的口语中，属于非正式表达。褒义为主，偶尔也带有“狠活、有点东西”的意思。这是近几年在互联网、尤其是游戏、编程、科技圈子里流行起来的新兴俚语用法。

be cooking表示“正在进步；做得非常好”，英文解释为“to be making progress or doing something well”

OpenAI said in a press release that Images 2.0 can also generate pictures from prompts in multiple languages, including Japanese, Korean, Chinese, Hindi, and Bengali.

OpenAI在一份新闻稿中表示，Images 2.0还能根据多种语言的提示词生成图片，涵盖日语、韩语、中文、印地语和孟加拉语等。

prompt /prɒmpt/ 作名词，1）表示“（计算机屏幕上的）提示符（显示计算机已经准备好接受指令）”，英文解释为“a sign on a computer screen that shows that the computer is ready to receive your instructions”

📍在ChatGPT火热的今天，prompt通常指的是给模型的输入，用来激发或引导模型生成特定的输出或回答。简而言之，prompt是指用户提供给语言模型的一段文字或命令，以便模型据此理解用户的需求并给出相应的响应或信息。例如，在与像ChatGPT这样的大语言模型交互时，用户提出的问题或指令就是一个“prompt”。模型会根据这个prompt的内容分析、处理并生成回答或执行相关的任务。Prompt对于大语言模型来说是非常关键的，因为它直接决定了模型如何理解和回应用户的需求。

📍prompt中文里有译为“提示词”，prompt engineering 就是“提示工程/提示词工程”

2）表示“（给演员的）提词，提白”，英文解释为“words that are spoken to an actor who has forgotten what he or she is going to say during the performance of a play”

3）作动词，表示“促使；导致；激起”，英文解释为“to make sb decide to do sth; to cause sth to happen”举个🌰：His speech prompted an angry outburst from a man in the crowd. 他的讲话激起了人群中一男子的愤怒。

📺美剧《斯巴达克斯：血与沙》（Spartacus: Blood and Sand）中的台词提到：One cannot but wonder what would prompt such an act 不知他们为何遭此不幸。

4）作动词，表示“（尤指）给（演员）提词”，英文解释为“to help someone, especially an actor, to remember what they were going to say or do”举个🌰：I forgot my line and had to be prompted. 我忘词了，只好让人提词。

In samples provided by OpenAI, the model created realistic advertisements, a photo of what appears to be a physical magazine, and multiple pages of manga.

在OpenAI提供的样本中，该模型生成了逼真的广告、一张看似实体杂志的照片，以及多页漫画。

manga /ˈmæŋɡə/ 从日语借用过来的词汇，表示“（源于日本的）漫画”，英文解释为“Manga is a style of drawing, originally from Japan, that is used in comic books with adult themes.”举个🌰：Known as manga, Japanese comic art has grown into a cult here. 日本漫画艺术在这里已经十分风靡。

A spokesperson for OpenAI said the company does not intend to copy or reproduce specific artwork and that the images are generated from patterns it has learned.

OpenAI的一位发言人表示，该公司无意抄袭或再现特定的艺术品，这些图像是根据模型学到的模式生成的。

Outside mimicry, the model allowed for some creative twists on real-life celebrities and images, like an 1980s glam version of legendary investor Warren Buffett and a blueprint of a Business Insider editor's dinner.

除了模仿之外，该模型还能对现实中的名人和图像进行一些创意改编，例如传奇投资者沃伦·巴菲特的1980年代魅力版，以及用蓝图风格呈现的《商业内幕》某位编辑的晚餐。

mimicry /ˈmɪm.ɪ.kri/ 表示“（为逗乐而）模仿别人或动物的声音或行为”，英文解释为“the act of copying the sounds or behaviour of a particular person or animal, often in order to make people laugh ”举个🌰：The mockingbird is known for its mimicry of other birds. 知更鸟以善于模仿其他鸟类而闻名。

twist /twɪst/ 作动词，表示“转动，旋转；缠绕，盘绕”，英文解释为“to turn something, especially repeatedly, or to turn or wrap one thing around another”举个🌰：The path twists and turns for over a mile. 这条小路弯弯曲曲，长达1英里多。

作名词，1）表示“扭动，转动”，英文解释为“an act of twisting something”举个🌰：She gave the cap another twist to make sure it was tight. 她又拧了一下盖子，以确保盖紧了。

2）表示“（故事或情况的）转折，转变，突然变化”，英文解释为“an unexpected change or development in a story or situation”举个🌰：The story has taken another twist. 故事情节再一次变化。

The spokesperson added that OpenAI prevents copying styles of individual living artists, but it does allow recreating broader studio styles.

这位发言人补充说，OpenAI禁止复制在世艺术家的个人风格，但允许再现更广泛的工作室风格。

Mitch Stoltz, the IP litigation director for the Electronic Frontier Foundation, told Business Insider in an interview that the latest model alone doesn't raise copyright issues, and that the distinction OpenAI makes between images generated from specific artwork and those generated from patterns the model learned is important.

电子前沿基金会的知识产权诉讼主任米奇·斯托尔茨在接受《商业内幕》采访时表示，仅凭最新模型本身不会引发版权问题，而OpenAI在区分从特定艺术品生成的图像和从模型学到的模式生成的图像方面，这一点至关重要。

litigation /ˌlɪt.ɪˈɡeɪ.ʃən/ 表示“诉讼，起诉”，英文解释为“the process of taking a case to a court of law so that a judgment can be made”举个🌰：The company has consistently denied responsibility, but it agreed to the settlement to avoid the expense of lengthy litigation. 公司一直在推卸责任，但同意这一调解方案以避免漫长的诉讼所带来的花费。

copyright /ˈkɒp.i.raɪt/ 表示“版权，著作权”，英文解释为“the legal right to control the production and selling of a book, play, film, photograph, or piece of music”举个🌰：Who owns/holds the copyright on this article? 谁拥有这篇文章的著作权？

📍单词“copyright”由词根“-right”（表示“权利”）和前缀“copy-”（表示“复制”）组成。想象你拥有这本书的复制权，这就是“copyright”。

"If the output is substantially similar to something that the model was trained on or crawled, then there starts to be a copyright concern," Stoltz said. "But if it's similar on the level of an idea, or just something that occurs in the world, or even a style or a vibe — generally speaking, that's not enough, and that's not a copy under copyright law."

“如果输出结果与模型训练或爬取的内容实质性相似，那么就会引发版权问题，”斯托尔茨说。“但如果相似是在想法层面，或者只是世界上存在的一些事物，甚至是一种风格或氛围——一般来说，这还不够，根据版权法这不构成抄袭。”

substantially /səbˈstæn.ʃəl.i/ 1）表示“大体上；基本上”，英文解释为“generally”举个🌰：This model has leather seats, but the two cars are substantially the same. 这一型号有几个附加装置，但两款车大体上是一样的。

2）表示“在很大程度上”，英文解释为“to a large degree”举个🌰：The new rules will substantially change how we do things. 新规则会在很大程度上改变我们的行事方式。

OpenAI is facing at least a dozen ongoing copyright suits from writers and news outlets, including The New York Times and George R.R. Martin.

OpenAI正面临来自作家和新闻媒体的至少十几起正在进行的版权诉讼，包括《纽约时报》和乔治·R·R·马丁。

"The copyright issues are the same as if you were simply using Photoshop, a darkroom, or a human artist to make these kinds of replicas," he said. "The issues for society, whether you call that ethical or political or whatever, are greater because it's easier, faster, and more available."

“版权问题与仅仅使用Photoshop、暗房（传统胶片修图技术）或人类艺术家制作这类复制品是一样的，”他说。“对社会而言的问题，无论你称之为道德、政治还是其他什么，都更严重，因为它更容易、更快、更易获得。”

- 词汇盘点 -

- 词汇助记 By AI -

LearnAndRecord

2015年2月8日

2026年4月26日

第4096天

每天持续行动学外语