ChatGPT 可以生成以假亂真的收據，有圖也沒真相？
You can use 4o to generate fake receipts.There are too many real world verification flows that rely on 'real images' as proof. That era is over. pic.twitter.com/9FORS1PWsb
— Deedy (@deedydas) March 29, 2025
由上面的例子可以看到，雖然說乍看之下好像有點不協調（主要是分隔線沒有隨著紙面的皺折而曲折，墨水的顏色也太均勻），但乍看之下好像還蠻逼真的。更何況，真正有心造假的人士可以很簡單地在影像編輯軟體中進行小幅度的修改，或甚至可能在下提示詞時更精準一點，就有可能讓它更加以假亂真了。下面就是另一位網友以前圖為基礎，加上了更多指示的結果，好像確實是又逼真了一點呢。
I think in the original image the letters are too perfect and they don't bend with the paper. They look like hovering above the paper. Here is my attempt to make it more realistic. Let me know what you think. pic.twitter.com/EixRSHubeY
— Michael Gofman (@michaelgofman) March 29, 2025
顯然，這讓「有圖有真相」又再次打上了一個巨大的問號，並且為不法份子「報銷」虛假的開銷提供了門路OpenAI 的發言人 Taya Christianson 向 TechCrunch 表示，他們所有的圖像都包含元數據（metadata），標明是由 ChatGPT 生成的。Christianson 並補充說，當用戶違反其使用政策時，OpenAI 會「採取行動」，並且他們會「持續從用戶在真實世界中的使用與反饋中學習」。
至於為什麼 ChatGPT 一開始就允許人們生成偽造收據，以及這是否符合 OpenAI 禁止詐欺的使用政策？對此Christianson 表示，OpenAI 的目標是「給予用戶盡可能多的創作自由」，而且假收據可能用於非詐欺的情境，例如教導人們金融知識，或是創作原創藝術和產品廣告等。
相關新聞：
ChatGPT 圖像生成功能已向免費用家開放，一日內最多 3 張
ChatGPT 單小時新增用家數破百萬，CEO 預告開源模型
ChatGPT 影像生成太火熱，Altman 表示 OpenAI 的 GPU「快融化了」
👉 變身吉卜力角色、吊牌公仔！實測 ChatGPT AI 最佳提示詞 + 平價訂閱會員方法
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Miami Herald
20 minutes ago
- Miami Herald
Microsoft exec offers horrifying advice to laid off employees
Microsoft (MSFT) made an announcement last week that's become hauntingly familiar news in recent months: it's decided to conduct yet another round of layoffs. On June 2, the Redmond-based tech company said it will lay off less than 4% of its global workforce, which adds up to about 9,000 of its employees. Don't miss the move: Subscribe to TheStreet's free daily newsletter Microsoft has done multiple rounds of layoffs this year as it scales up its AI plans, leaving many wondering if they'll be next on the chopping block. It's a scary time to be in tech, as Meta, Amazon, and Google have also announced layoffs this year. All these companies have also been building their AI investments, with Meta committing to spend $65 billion in 2025, Amazon committing $100 billion, and Google committing to $75 billion. Related: Microsoft sends a brutal message to loyal employees Add into this that the news is crammed with threats of AI eliminating jobs, and it's naturally creating a lot of anxiety for those who have been laid off and need to look for a new role. People often wax philosophical on LinkedIn about layoffs, whether they were affected by them or simply want to comment on what they mean for the affected industry. But when a key Microsoft executive decided to do so in a recent post, he hit a nerve with comments that seemed tone-deaf to the newly unemployed. In a now-deleted LinkedIn post screencapped by BlueSky user and Necrosoft Games Creative Director Brandon Sheffield, Executive Producer at Xbox Game Studios Publishing Matt Turnbull shared some thoughts on what he believed those affected by the recent Microsoft layoffs should do. More Layoffs: Disney makes a devastating layoffs announcementFedEx layoffs signal a concerning business trendMeta quietly plans rude awakening for employees after layoffs "These are really challenging times, and if you're navigating a layoff or even quietly preparing for one, you're not alone, and you don't have to go it alone," Turnbull wrote. "I know these types of tools engender strong feelings in people, but I'd be remiss in not trying to offer the best advice I can under the circumstances," he continued. "I've been experimenting with ways to use LLM AI tools (like ChatGPT or Copilot) to help reduce the emotional and cognitive load that comes with job loss." Related: Microsoft joins Facebook and Google in upsetting practice Turnbull goes on to recommend a few prompts those affected by layoffs can use, including career planning prompts, resume help, how to draft a post to foster networking, and even how to ask an LLM for help with imposter syndrome. In his own post where he shared the screencap, Sheffield said, "Something I've realized over time is people in general lack the ability to think in a broader scope and include context and eventualities. But after thousands of people get laid off from your company, maybe don't suggest they turn to the thing you're trying to replace them with for solace." While the June jobs report looked positive at a glance, with the unemployment rate falling to 4.1% and 147,000 nonfarm jobs added per the Bureau of Labor Statistics, a deeper look reveals other troubles. Unemployment duration increased from 21.8 to 23 weeks, signifying a longer struggle to find work, and the percentage of people unemployed for 27 weeks or more rose to 23.3%. More stories are also emerging about the use of AI in both hiring and applying. Jobseekers have begun to report experiencing job interviews with AI instead of with human beings, with many finding the process clinical and disappointing. To fight back against what seems like an impossible job market, many applicants are also using AI to mass apply to jobs, leaving recruiters overwhelmed and unable to cope. Meanwhile, a recent ResumeBuilder survey also found that 66% of managers now turn to AI chatbots such as ChatGPT when making decisions about layoffs. In addition, 78% turn to it for decisions on raises, while 77% use it for deciding on promotions. Related: Amid AI boom, veteran analyst reboots AMD, Supermicro stock price targets The Arena Media Brands, LLC THESTREET is a registered trademark of TheStreet, Inc.
Yahoo
2 hours ago
- Yahoo
OpenAI's Wallet Just Got Heavier -- Because Meta's Been Poaching
OpenAI isn't just building cutting-edge AI anymore; it's now fighting a full-blown talent war and it's using stock options as its shield. The company known worldwide for ChatGPT has seen its stock-based compensation jump more than five times over the past year; that's not a typo. In total, OpenAI handed out $4.4 billion in equity; that figure was 119% of its entire revenue for the same period. Yes it's now literally paying out more in stock than it's earning. Warning! GuruFocus has detected 6 Warning Sign with META. And the trigger? Meta (META, Financials) which has reportedly lured away at least nine researchers from OpenAI's AI team including some working on foundational models. These exits weren't minor; they cut deep. OpenAI had hoped this stock-based spending spree would cool off by 2025 projecting equity payouts to drop to 45% of revenue, and then under 10% by the end of the decade. But that was before Meta started raiding its brain trust. Now, those assumptions are out the window. According to internal chatter, Chief Research Officer Mark Chen believes the company might have to sweeten its equity offers even more; because when your biggest assets walk out the door, the only thing left to do is open it wider with better incentives. This isn't just about keeping salaries competitive; it's about survival. The AI space is getting fiercer by the week; and even with Microsoft in its corner, OpenAI can't afford to lose talent to rivals. Equity compensation isn't a perk anymore; it's a weapon. And OpenAI is loading up. This article first appeared on GuruFocus.
TechCrunch
2 hours ago
- TechCrunch
Cursor apologizes for unclear pricing changes that upset users
The CEO of Anysphere, the company behind the popular AI-powered coding environment Cursor, apologized Friday for a poorly communicated pricing change to its $20-per-month Pro plan. The changes resulted in some users complaining that they unexpectedly faced additional costs. 'We recognize that we didn't handle this pricing rollout well and we're sorry,' said Anysphere CEO Michael Truell in a blog post. 'Our communication was not clear enough and came as a surprise to many of you.' Truell is referring to a June 16 update to Cursor's Pro plan. Instead of Pro users getting 500 fast responses on advanced AI models from OpenAI, Anthropic, and Google, and then unlimited responses at a slower rate, the company announced subscribers would now get $20 worth of usage per month, billed at API rates. The new plan allows users to run coding tasks in Cursor with their AI model of choice until they hit the $20 limit, and then users have to purchase additional credits to continue using it. However, Pro users took to social media to file their complaints in the weeks following the announcement. Many users said they ran out of requests in Cursor rather quickly under the new plan, in some cases after just a few prompts when using Anthropic's new Claude models, which are particularly popular for coding. Other users claimed they were unexpectedly charged additional costs, not fully understanding they'd be charged extra if they ran over the $20 usage limit and had not set a spend limit. In the new plan, only Cursor's 'auto mode,' which routes to AI models based on capacity, offers unlimited usage for Pro users. Anysphere says it plans to refund users that were unexpectedly charged, and aims to be more clear about pricing changes moving forward. The company declined TechCrunch's request for comment beyond the blog post. Truell notes in the blog that Anysphere changed Cursor's pricing because 'new models can spend more tokens per request on longer-horizon tasks' — meaning that some of the latest AI models have become more expensive, spending a lot of time and computational resources to complete complicated, multi-step tasks. Cursor was eating those costs under its old Pro plan, but now, it's passing them along to users. Techcrunch event Save $450 on your TechCrunch All Stage pass Build smarter. Scale faster. Connect deeper. Join visionaries from Precursor Ventures, NEA, Index Ventures, Underscore VC, and beyond for a day packed with strategies, workshops, and meaningful connections. Save $450 on your TechCrunch All Stage pass Build smarter. Scale faster. Connect deeper. Join visionaries from Precursor Ventures, NEA, Index Ventures, Underscore VC, and beyond for a day packed with strategies, workshops, and meaningful connections. Boston, MA | REGISTER NOW While many AI models have lowered in price, the cutting edge of performance continues to be expensive — in some cases, more pricey than ever. Anthropic's recently launched Claude Opus 4 model is $15 per million input tokens (roughly 750,000 words, longer than the entire 'Lord of The Rings' series) and $75 per million output tokens. That's even more costly than Google's launch of Gemini 2.5 Pro in April, which was its most expensive AI model ever. In recent months, OpenAI and Anthropic have also started charging enterprise customers for 'priority' access to AI models — an additional premium on top of what AI models already cost that guarantees reliable, high speed performance. These expenses may be filtering their way down to AI coding tools, which seem to be getting more expensive across the industry. Users of another popular AI tool, Replit, were also caught off guard in recent weeks by pricing changes that made completing large tasks with AI more expensive. Cursor has become one of the most successful AI products on the market, reaching more than $500 million in ARR largely through subscriptions to its Pro plan. However, Cursor now faces intense competition from the AI providers it relies on, while simultaneously figuring out how to affordably serve their more expensive AI models. Anthropic's recently launched AI coding tool Claude Code has been a hit with enterprises, reportedly boosting the company's ARR to $4 billion, and likely taking some users from Cursor in the process. Last week, Cursor returned the favor by recruiting two Anthropic employees that led product development of Claude Code. But if Cursor intends to keep its market-leading position, it can't stop working with the state-of-the-art model providers — at least, not until its own home-grown models are more reasonably competitive. So Anysphere recently struck multi-year deals with OpenAI, Anthropic, Google, and xAI to offer a $200-a-month Cursor Ultra plan with very high rate limits. Anthropic co-founder Jared Kaplan also told TechCrunch in June he plans to work with Cursor for a long time. However, it certainly feels as if the pressure between Cursor and AI model developers is building.
