Author: admin

• Anthropic’s push to label AI-generated text is prompting some to create tools to remove those labels.
• One AI watermark remover has gone viral on GitHub, and search interest for the tools has increased.
• AI watermarks have divided users, and some have canceled their Claude subscriptions.

Anthropic wants to make AI-generated text easier to identify. Some software developers are already building tools designed to make those labels disappear.

The AI company said last week the “imperceptible watermark” on text produced by certain Claude models would travel with it when copied and pasted because it’s a statistical pattern created through Claude’s word choices.

While some have welcomed the improved transparency that AI watermarks could offer, the move has also prompted concerns among techies that a label could follow work even if they only used Claude to proofread, translate, or summarize. Some users previously told Business Insider they have canceled their Claude subscriptions because of it.

The backlash has led others to seek out ways to get around it: Google Trends interest for “AI watermark remover” in the US is up 60% week-on-week. Some developers and entrepreneurs have taken matters into their own hands and created tools to remove AI watermarks from text and files.

Anthropic has said it’s introducing the watermark to meet its commitments to the European Union’s AI Act. The company did not respond to Business Insider’s questions about AI watermark removal tools.

‘The wrong answer to a real problem’

Guillaume Meyer, a Paris-based tech entrepreneur and the founder of Memo, an AI tool for e-commerce brands, released an open-source project called “Watermarks Remover” days after Anthropic announced its plans.

The tool strips hidden characters and metadata, then rewrites text to preserve its meaning, thereby disrupting the statistical word-choice patterns that can carry a watermark, Meyer said.

The first version took about five hours to build, he added. It has more than 14,000 GitHub stars, a measure of its popularity among developers that’s similar to “likes” on social media sites. It does not guarantee watermark removal.

“I am all for content attribution,” Meyer told Business Insider. “I am against the watermarking technique, and that’s a very significant distinction.”

Meyer said his objection is that watermarking “treats authorship as a binary thing.” AI could leave a watermark when used to generate the text entirely, but also for lighter editing.

Anthropic said in a blog post that its watermark is not meant to establish authorship.

“I think it’s the wrong answer to a real problem,” Meyer said of efforts to watermark AI-generated text.

A new breed of AI watermark removers

Meyer is not alone in trying to take on AI watermarks.

Sabrina Ramonov, an AI educator for entrepreneurs, said in an X post this week that she created a free, browser-based “Watermark Remover” that says it can “clean hidden AI marks” from text, PDFs, Word documents, web pages, images, and data files.

Meanwhile, Ansh Aneja, a Tokyo-based software developer, said he built a Claude-focused watermark remover on the day Anthropic announced the feature. He later released a local, open-source version called MarkScrub.

In an X post last week, Aneja said the tool went from zero to 8,500 users in a day. Business Insider could not independently verify those figures.

These projects are still nascent, and Business Insider has not been able to independently assess their effectiveness. Other companies, including Google and OpenAI, also use watermarks for images, and there are scores of tools that promise to remove them.

Still, the tools illustrate a long-standing limit of watermarking, said Thibaud Gloaguen, a researcher at ETH Zurich’s Secure, Reliable, and Intelligent Systems Lab.

“There will always be ways to remove the watermark,” he told Business Insider, giving the example of rewording the entire text.

A gray area

Watermark removers sit in a legal gray area: the EU tells AI companies to add durable labels, but does not clearly set out rules for third parties seeking to remove them.

A European Commission spokesperson told Business Insider the bloc’s guidance requires providers’ marking systems to be resilient to common alterations and adversarial actions against them.

The spokesperson added the EU’s transparency code specifically identifies removal, regeneration, copying, and modification as threats that providers should assess.

Those are requirements for AI providers, however. The AI Act does not expressly ban third parties from trying to remove a watermark, said Dmitri Roussinov, a senior lecturer in Computer and Information Sciences at the University of Strathclyde.

If that same removal tool uses AI to regenerate text, though, its provider may have its own obligation under EU law to mark that newly AI-generated output, he told Business Insider.

Konrad Kollnig, an assistant professor at Maastricht University’s Law and Tech Lab, said neither the EU’s AI Act nor Anthropic’s terms appear to bar people from making or sharing removal tools.

Users could, however, face problems if they use one to misrepresent AI-generated work as human-made. Anthropic’s usage policy prohibits impersonating a human by presenting model results as human-generated.

That question could become more urgent when Anthropic releases its planned text-detection API alongside its next model. It has not specified a date yet. Watermarks will apply to new models released on or after August 2, and Anthropic is also planning to roll it out on older models.

“All these watermarking techniques have one big issue,” Kollnig said. “Whenever the watermarking detection tool is made public, anyone can check AI-generated contents — including text and videos — against that detector and can build tools to remove watermarks.”

Are you using an AI watermark remover at work? Contact this reporter via email at tspirlet@businessinsider.com or Signal at thibaultspirlet.40. Use a personal email address, a nonwork WiFi network, and a nonwork device; here’s our guide to sharing information securely.

• China reacted to US import bans by targeting seven American firms, including Altana.
• Altana’s software helps companies manage visibility and compliance of their complex supply chains.
• Wayfair collaborates with Altana’s AI to trace supply chain risks and improve regulatory adherence.

Altana, a software company that helps brands like Wayfair map their global supply chains, has been barred by the Chinese Ministry of Commerce from doing business with individuals or companies in the country.

It’s an example of how geopolitics continue to complicate the supply chain, even while technology advancements like AI offer new solutions.

The August 5 move came after the US banned imports from dozens of additional Chinese companies over alleged human rights abuses in late July 2026. In addition to Altana, the six other companies on China’s list are the nonprofits Responsible Business Alliance and Verité Group, the human-rights NGO Human Rights in China, and the natural-fiber lab-testing company Stratum Reservoir. The countermeasures list also included Applied DNA Sciences, which recently changed its name to BNB Plus Corp and pivoted from fiber tracking to blockchain services in the cryptocurrency space.

Altana sells its software to businesses so they can uncover forced labor risks and meet strict regulatory compliance. To do that, Altana collects and buys data points from businesses around the world, including in China.

The company’s inclusion on the list could make it more difficult to provide some of this data to its clients, like Wayfair.

A Wayfair spokesperson told Business Insider, “While this is a new development, we remain committed to using technology to see deeper into and better understand our supply chain.”

Altana declined to provide a comment to Business Insider about the actions and how they will affect its software going forward, but did previously speak to Business Insider about the work it has been doing on supply chain mapping.

Demystifying hidden supply chain layers

For a company like Wayfair — one that sources 40 million products globally from 20,000 suppliers — abiding by US government prohibitions requires continuous visibility and cross-checking across multiple supply chain tiers.

It’s a complex process since the company “can’t manage what we don’t see,” Evan Friedler, the director of regulatory and government affairs at Wayfair, told Business Insider during a July interview.

Previously, Wayfair contracted with third-party companies and used internal systems to cross-check its supply chain against unauthorized parties, but “getting to the first tier of the supply chain has historically been the easy part,” Friedler said. Visibility often ended there, leaving factors like questionable labor practices or prohibited companies working in second or third-tier suppliers a mystery. Wayfair’s increasingly global scale and a rising regulatory bar made it more important to reach those other tiers, and AI technology finally made it more possible, Friedler said.

About two years ago, Wayfair began working with Altana to go beyond that first tier and look higher up on the supply chain — and with fewer manual checks, Friedler said. “This solution helps us proactively manage risk rather than react to it.”

AI data processing with a built-in paper trail

Evan Smith, Altana’s cofounder and CEO, said the platform is like “Google Maps for the supply chain.”

Altana draws from paid and free public-domain data and secures rights to certain insights from client and user data sets as they are processed through its system, said Smith. The company won’t share confidential customer data, such as pricing or bill-of-materials details.

Still, they can use the data they obtain to enrich the existing picture of the world supply chain graph, said Smith. “It’s a contributory network where everyone’s hooking their data and their users into a platform that allows for that shared visibility,” he said.

At Wayfair, Altana is used for labor compliance and denied-party issues. For example, Altana notifies Wayfair when suppliers potentially match a government-published list of forced labor or denied parties. Wayfair’s compliance team can investigate these risks sooner than before it adopted the technology, said Friedler. Previously, Wayfair used manual signals and self-reported supplier information for its supply chain diligence.

Altana also uses agentic AI workflows to ingest and store vast quantities of data, pulling in reference information such as the full body of US Customs rulings on product classification. Wayfair workers can click into the provided background information to see the AI’s reasoning and underlying documentation for potential issues so they have the evidence they need to substantiate any AI-issued determinations.

Most of Wayfair’s alerts are ultimately false positives. The Wayfair compliance team investigates potential issues, which can mean confirming company names and spellings, double-checking the supplier’s region, or calling suppliers for more information.

Wayfair then updates the Altana alerts on the platform with this information, which they’d typically be unable to uncover through their own manual processes, Friedler said.

• McDonald’s is losing sales among low-income diners, new data shows.
• Numerator found that sales among low-income customers slipped 2.4% duing the chain’s latest quarter.
• McDonald’s has said that the execution of its latest value menu hampered second-quarter sales.

McDonald’s is losing ground to rivals among low-income fast-food diners, according to recent data.

Spending by low-income guests at the Golden Arches declined 2.4% year-over-year during the company’s latest quarter, said consumer analytics company Numerator. The firm defines “low-income guests” as those from households making $40,000 or less a year.

The drop is McDonald’s first quarterly decline with low-income guests in the past year and amounts to roughly $310 million in lost sales, Numerator said on Tuesday.

Rival fast-food chain Burger King, meanwhile, notched a 0.3% gain over the same period.

The data is the latest sign that McDonald’s is struggling to attract diners, especially in a K-shaped economy.

While the chain long won over price-conscious diners with deals like its buy-one-get-one discount on sandwiches such as the Big Mac, it has had issues rolling out its latest value menu, which focuses on items priced $3 or less, CEO Chris Kempczinski said on an earnings call earlier this month.

“Although we’ve restored our overall value and affordability leadership, our restaurant-level results show that execution was inconsistent across the system,” Kempczinski said.

McDonald’s comparable US sales growth slowed to 0.8% during its second quarter, the company said. Rivals from Burger King to casual dining chain Chili’s have posted stronger results while offering their own value menus and deals, Business Insider has reported.

Some McDonald’s customers have told Business Insider they’re now eating more at other chains or cooking more at home, due to higher costs.

Despite the company’s struggles, some McDonald’s offers are still drawing in diners of all income levels, Numerator found.

The chain’s fried apple pie, an old menu item it brought back for America’s 250th birthday this summer, has been purchased by 11.7% of US households, Numerator said.

Do you have a story idea about McDonald‘s? Contact this reporter at abitter@businessinsider.com or via encrypted messaging app Signal at 808-854-4501. Use a personal email address, a nonwork WiFi network, and a nonwork device; here’s our guide to sharing information securely.

• Pennsylvania Gov. Josh Shapiro announced a crackdown on data centers.
• Shapiro said that incomplete and lackluster proposals are overwhelming local officials.
• He said the speculators are “scaring our communities.”

Pennsylvania Gov. Josh Shapiro is cracking down on data centers, becoming the latest state leader to scale back support for the massive facilities critical to the future of AI.

Shapiro painted a stark picture of communities being overrun by half-baked data center proposals, even as he stressed that only five such facilities have the permits they need to begin operating.

“These speculators are nevertheless scaring our communities — being aggressive with township officials, bullying our neighbors, and refusing to listen to the people of Pennsylvania,” Shapiro said during an announcement Tuesday in Harrisburg. “And they’re threatening to fundamentally change the character of our communities.”

It’s unclear what the announcement means for Amazon’s planned $20 billion investment for data center campuses in Bucks and Luzerne counties. A spokesperson for Amazon did not immediately respond to a request for comment from Business Insider.

In Pennsylvania and across the country, that AI future has run into a buzzsaw of public dissent. Now, governors in both parties are pushing their states to rein in incentives for data center development. Last month, New York Gov. Kathy Hochul imposed the nation’s first statewide moratorium.

Under the order, Shapiro said he was immediately removing data centers from fast-track permitting. Most critically, the order will also offer local communities an effective veto over a data center project’s permits.

“Let me put it succinctly: if the local community doesn’t approve a project, the state won’t approve it either,” he said.

Shapiro, who is widely viewed as a 2028 presidential hopeful, said that local opposition to proposed data centers had shaped his view on the topic. He also said that the legislature’s failure to act “left him with no other choice” but to act on his own.

Shapiro called out projects by name that are unrelated to the Amazon plans. In describing one proposal, Shapiro lamented that a developer wants to build on top of local baseball fields.

“That left field line where parents should gather to watch their kids play Little League,” he said. “Instead, that’s where the corner of one of those data centers was to be built.”

“I want you to know I hear you,” he said.

• Eric Schmidt is selling his superyacht Whisper.
• The luxury ship, which he bought three years ago, is listed for $172.5 million.
• Originally built for billionaire Shahid Khan, Whisper features a spa, gym, and movie theater.

If you’re in the market for a shiny new toy, Eric Schmidt’s superyacht is for sale.

Whisper, which the billionaire former Google CEO purchased from Shahid Khan in 2023, is listed for $172.5 million.

Built by the German shipyard Lürssen, which also constructed David Geffen’s Rising Sun, Whisper measures 95 meters long. Spread over five decks, it features a master cabin with a fireplace, a spa with a cold plunge and steam room, a movie theater, and a swimming pool. It comes equipped with Jet Skis, wakeboards, and kayaks.

The megayacht fits 12 guests and a 27-person crew — a comfortable ratio — and can entertain many more, with its charter brochure boasting that “Whisper can — and has — hosted the largest parties.”

“Fusing high glamour with a guest-centric, service-oriented layout, Whisper’s five larger-than-life decks contain every amenity you could possibly need for a fun-packed week at sea,” it says.

Originally built for Jacksonville Jaguar-owner Khan, the yacht was delivered as Kismet in 2014. It was listed for $161 million when Schmidt purchased it three years ago, and a retrofit was completed last year.

There is no information about why Schmidt, who is worth $58.2 billion, according to Bloomberg, is selling the yacht, and his representative declined to comment to Business Insider.

Perhaps he’s looking to take advantage of the growing appetite for superyachts, a result of the surging market and the AI boom, or perhaps it’s that Whisper was never Schmidt’s original plan.

In 2023, he nearly purchased Alfa Nero, an 81-meter megayacht seized from a sanctioned Russian oligarch, though he backed out amid legal challenges, Bloomberg reported at the time.

There’s some good news for those who don’t have nine figures to spare. Whisper is also available to charter for a comparative bargain of $1.4 million a week.

• Espresso martinis are one of my favorite cocktails, so I decided to try Ina Garten’s version. 
• Garten uses orange-flavored vodka, along with fresh espresso and Kahlúa. 
• Her espresso martini definitely packs a punch. I was buzzed after a few sips!

Ina Garten famously loves a good cosmopolitan, but there are plenty of other cocktail recipes to be found in her cookbooks.

The “Barefoot Contessa” star was even ahead of the curve when it came to espresso martinis, demonstrating how to make them on a 2007 episode of her Food Network show.

Now you’ll see espresso martinis on just about every bar menu, but they don’t come cheap (at least in LA or NYC). So when I stumbled on Garten’s recipe, I knew I had to take it for a spin.

Garten’s espresso martini recipe has only three ingredients.

All you need to whip up Garten’s drink is: 

• 2 cups of freshly brewed espresso, chilled
• 2 cups of orange-flavored vodka
• 1 cup of coffee liqueur (Garten recommends Kahlúa)

These measurements make eight cocktails total. If you want to mix individual drinks, Garten recommends using 2 ounces of espresso, 2 ounces of orange vodka, and 1 ounce of coffee liqueur per person.

To get started, I brewed some espresso.

Garten notes in her recipe that the coffee for your espresso martini should be “freshly brewed.” 

“I like to make it with decaf,” she explained during an episode of her “Barefoot Contessa” cooking show. “I don’t want to have jittery guests.” 

Since I was making my cocktail as the afternoon slump was setting in, I decided to go with a caffeinated version instead. After brewing the espresso, I popped it into the fridge for 30 minutes to chill.

After my espresso had chilled, I began building the cocktail.

When I first tried making the cocktail, I didn’t realize Garten throws all of her ingredients into a big pitcher first. The step isn’t noted in the Food Network recipe, so I assumed I was supposed to put everything straight into the shaker. 

My shaker was already almost full when I added the vodka, a clear sign that I had messed up.

At first, I figured my cocktail shaker was just too small for the recipe. I tried to dump some ice cubes out to make room for the Kahlúa, but it didn’t make much of a difference. 

As I poured the Kahlúa into the shaker, I knew something was definitely off. I had already halved the measurements to make four drinks instead of eight, but there still wasn’t enough room in the shaker.

I dumped a little more out before pouring in the rest of the Kahlúa and hoped for the best.

And when I poured the drinks, it was obvious that something had gone wrong.

My cocktails didn’t have the signature foamy top that you always see on an espresso martini. But maybe it would taste OK? 

It did not.

The alcohol was overwhelming, and not in a fun way. To be honest, it might have been one of the worst drinks I’ve ever made. And I used to chase shots of vodka with water in college!

I realized the cocktail wasn’t able to mix or foam because there was almost no space left in the shaker. I had to go back to the drawing board.

I found the episode where Garten makes her espresso martini and realized I had missed a crucial step.

This is when I realized Garten puts all of her ingredients into a big pitcher and mixes them together. When it’s time to serve, she mixes each individual drink in the shaker. 

Oh, duh. 

In retrospect, it seems obvious, as that’s exactly what Garten did for her big batch of famous cosmos. But I thought I was following the recipe! 

Anyway, we live and learn. Time for round two. 

Second time was most definitely the charm. My cocktail came out looking like a classic espresso martini.

The top had a nice layer of foam, allowing the little coffee beans to pop. Once I added the orange peel twist, I couldn’t help but feel proud of myself. This drink looked legit. 

The vodka taste wasn’t as overpowering as it had been in the first drink, but it was still strong. If you’re not a huge fan, I’d make the ratio of vodka equal to the Kahlúa.

What surprised me the most about Garten’s espresso martini was how strong it is. I’m no lightweight, but I was buzzed after a few sips!

If you’re looking for a drink that will wake you up rather than knock you off your feet, you might want to cut the alcohol down a bit.

But if you’re hoping for a big night out, Garten’s espresso martini will definitely kick things off.