Elon Musk's xAI recently weakened content guard rails for image generation in the Grok AI bot. This led to a new spate of non-consensual sexual imagery on X, much of it aimed at silencing women on the platform. This, along with the creation of sexualized images of children in the more compliant Grok, has led regulators to begin investigating xAI. In the meantime, Google has rules in place for exactly this eventuality—it's just not enforcing them.

It really could not be more clear from Google's publicly available policies that Grok should have been banned yesterday. And yet, it remains in the Play Store. Not only that—it enjoys a T for Teen rating, one notch below the M-rated X app. Apple also still offers the Grok app on its platform, but its rules actually leave more wiggle room.

App content restrictions at Apple and Google have evolved in very different ways. From the start, Apple has been prone to removing apps on a whim, so developers have come to expect that Apple's guidelines may not mention every possible eventuality. As Google has shifted from a laissez-faire attitude to more hard-nosed control of the Play Store, it has progressively piled on clarifications in the content policy. As a result, Google's rules are spelled out in no uncertain terms, and Grok runs afoul of them.

Google has a dedicated support page that explains how to interpret its "Inappropriate Content" policy for the Play Store. Like Apple, the rules begin with a ban on apps that contain or promote sexual content including, but not limited to, pornography. That's where Apple stops, but Google goes on to list more types of content and experiences that it considers against the rules.

"We don’t allow apps that contain or promote content associated with sexually predatory behavior, or distribute non-consensual sexual content," the Play Store policy reads (emphasis ours). So the policy is taking aim at apps like Grok, but this line on its own could be read as focused on apps featuring "real" sexual content. However, Google is very thorough and has helpfully explained that this rule covers AI.

The detailed policy includes examples of content that violate this rule, which include much of what you'd expect—nothing lewd or profane, no escort services, and no illegal sexual themes. After a spate of rudimentary "nudify" apps in 2020 and 2021, Google added language to this page clarifying that "apps that claim to undress people" are not allowed in Google Play. In 2023, as the AI boom got underway, Google added another line to note that it also would remove apps that contained "non-consensual sexual content created via deepfake or similar technology."

Sound like any apps you know?

The archetype of a bannable app, approved for teens

Taken together, Google's description of bannable apps describes Grok's app to a tee. Google made these additions as new threats became apparent, knowing that developers would try to publish AI-undressing apps in the Play Store. The company did not, apparently, think the world's richest person would be the one pushing digital humiliation tools on its platform. And Google's response to this situation so far has been to do nothing.

The backlash to xAI's loosened restrictions prompted the company to limit access to image editing slightly. You can no longer edit images on X without paying for a premium plan. However, the Grok app does not have that limitation. Anyone who downloads Grok can use it to create non-consensual sexual content.

Since the app is cleared for teens, even devices with parental controls enabled will permit 13- to 17-year-olds to download Grok. There is no paywall, and you don't even have to log in before editing your first image. The app does ask the user to confirm their birth year, but teenagers would never lie about that, right?

This is not xAI's first problem with non-consensual sexual content. Last year, the AI was widely used to create fake Taylor Swift nudes. However, in that case, users were simply prompting the bot with the singer's name—Grok can create entirely new images of famous people because the training data includes real images of them. Grok's newer ability to "edit" images of people is a different and more insidious feature because it can turn anyone into an AI plaything.

Ars has reached out to Google to ask why Grok has not been removed and why it has retained a Teen rating. The company has declined to make a statement at this time. So we're left with a policy that explicitly bans apps like Grok, but Google is taking no action to enforce those policies, allowing impressionable teenagers and unsavory weirdos to use it to sexualize real people.

Read full article

Comments

“Kristi Noem’s an idiot, right? Like, she went on TV before any of the facts were in. She made up this story on TV about ICE agents trying to dig their truck out of the snow or something, which was clearly false. Like she didn’t even realize that there was gonna be video of what happened when she went up and gave a false story.”

— Rep. Seth Magaziner (D-RI), quoted by Mediaite.

Collecting fallen (or "shed") elk antlers is a popular pastime in elk-heavy places like Montana, but it's usually a pretty low-tech, on-the-ground affair. That's why last year's story about a US Black Hawk helicopter descending from the skies to harvest shed elk antlers on a ranch was such an odd one.

Was it really possible that US military personnel were using multimillion-dollar government aircraft to land on private property in the Crazy Mountains—yes, that's their actual name—just to grab some antlers valued at a few hundred bucks?

Antler hunt

In May 2025, Montana rancher Linda McMullen received a call from a neighbor. "He said, 'Linda, there’s a green Army helicopter landed on your place, picking up elk antlers,’” McMullen told The New York Times last year. “I said, ‘Are you joking?’ He said, ‘I’m looking at them with binoculars.’”

The local sheriff, who said he was "still trying to figure all this out" at the time, added that this was "the first helicopter I've heard of" regarding shed antler collection.

The Adjutant General for the Montana National Guard, J. Peter Hronek, quickly issued a statement on Facebook "regarding unauthorized use of military aircraft."

In it, Hronek said that he was "aware of an alleged incident involving a Montana Army National Guard helicopter landing on private property without authorization" and that "an internal investigation is underway, and appropriate adverse and/or administrative action will take place if the allegations are determined to be true." The Black Hawk was apparently on a training flight at the time.

The three servicemen on the chopper were eventually charged in Sweet Grass County Court with trespassing. They all pleaded not guilty. This week, pilot Deni Draper changed his plea to "no contest," allowing sentencing to go forward without a trial (but without actually admitting guilt).

According to local reporting, prosecutors had evidence that "no trespassing signs were posted on McMullen's property" and that "Draper admitted to Montana game warden Austin Kassner that he piloted the helicopter and decided to land it." In addition to the neighbor's testimony, "helicopter tire indentations and exhaust marks in the grass" were present at the site of the alleged landing.

The judge has accepted the change of plea and hit Draper with a $500 fine—the maximum penalty. So long as Draper stays out of trouble for the next six months, he will avoid further fines and jail time.

As for the antlers themselves, they are currently held by Montana Fish, Wildlife, and Parks but could go back to McMullen once cases against the other two servicemembers are resolved.

Read full article

Comments

Recall that Microspeak is not merely for jargon exclusive to Microsoft, but it’s jargon that you need to know to survive at Microsoft.

The term big rocks was introduced by Stephen Covey in the book The Seven Habits of Highly Effective People, which I suspect is very popular among senior executives, because senior executives aspire to become highly effective people.

In its original formulation, the concept of big rocks was used as a metaphor for time management: the metaphor is that you have a jar with large rocks inside it, stacked up to the brim. Is the jar full? But you can pour pebbles and sand into the jar to fill the gaps between the big rocks. The lesson is that you were able to fit everything into the jar if you put the big rocks in first. If you had started with the pebbles and sand, then there wouldn’t be space for the rocks. In terms of time management, the lesson is to deal with the biggest, most important things (the big rocks) first. If you spend time on the smaller things, you will find that there’s no room for the big things.

However, that’s not always what it means at Microsoft.

As I look over various types of documents, the meaning of big rocks as top priorities tends to predominate in senior executive documents.

These are the Big Rock priorities that have been determined by senior leadership.

And I was fortunate to find a document that opened with a definition.

The Nosebleed Big Rocks are the top business critical programs in our division.

However, as you go lower in the hierarchy and interact with people who do not keep a copy of The Seven Habits of Highly Effective People on their nightstand, the term big rocks tends to be used to mean the big problems that need to be solved in order for the project to succeed.

Again, I was able to find a document that included a definition.

Big Rocks: A list of technical challenges that we need to solve.

Bonus chatter: My theory (which has yet to be well-tested) is that if a speaker uses the term big rocks in a presentation, you can tell which definition the speaker is using by looking at the clip art they put on the slide. If it’s a bunch of boulders, then they use it to mean that it’s a problem to be solved. If it’s a jar, then they use it to mean a priority goal.

Narrator: It’s never a jar.

The post Microspeak: Big rocks appeared first on The Old New Thing.

In 2002, Bryan Fleming helped to create pcTattletale, software for monitoring phone and computer usage. Fleming's tool would record everything done on the target device, and the videos would be uploaded to a server where they could be viewed by the pcTattletale subscriber.

This might sound creepy, but it can also be legal when used by a parent monitoring their child or an employee monitoring their workers. These are exactly the use cases that were once outlined on pcTattletale's website, where the software was said to have "helped tens of thousands of parents stop their daughters from meeting up with pedophiles." Businesses can "track productivity, theft, lost hours, and more." Even "police departments use it for investigating."

But this week, nearly 25 years after launching pcTattletale, Fleming pled guilty in federal court to having knowingly built and marketed software to spy on other adults without their consent. In other words, pcTattletale was often used to spy on romantic partners without their knowledge—and Fleming helped people do it.

When you're sleeping

It's unclear when pcTattletale began marketing itself as a tool for catching cheaters, but Fleming's original business partner left the company in 2011, and Fleming ran things himself from his home in a northern Detroit suburb.

In 2021, Vice reported that pcTattletale was leaking the sensitive data it collected. The story quoted marketing materials about using the tool to catch a "cheating spouse," which required users to know their spouse's "pass-code and have access to the phone for about 5 minutes. The best time to do this is when they are sleeping." The company also provided instructions to hide icons that might reveal that pcTattletale was running on the victim's phone.

A look through archived versions of the pcTattletale site on the Wayback Machine shows that by 2022, pcTattletale had added numerous "cheating" links to its footers and featured multiple blog posts on ways to "catch your boyfriend cheating." These explicitly directed people to use the "unlock code to your boyfriend's phone" to install "the pcTattletale spy app" in order to "watch everything he does on his phone." One entry even noted that people being spied on in this way are unlikely to be happy about it, and users should "expect him to lash back at you over putting the spy app on his phone. It can really turn the tables."

Around this same time, federal investigators in California had launched an investigation into "stalkerware," and pcTattletale was among their targets. It also looked like a site where an arrest might not be too difficult, since Fleming operated out of the US and made no real attempts to hide his location. (Indeed, older versions of the pcTattletale website said explicitly that "Fleming Technologies" was based in Bruce Township, Michigan.) As a government investigator put it, "many of the other [stalkerware] websites under investigation involve targets who are believed to be overseas. For this reason, it is unrealistic to believe that the targets will soon be apprehended."

But Fleming was easy enough to find, and investigators soon obtained copies of his email account. It contained plenty of support requests in this vein: "Also if there is a way to NOT let user know you are taking screen shot that would be helpful too. My husband knows when there is screen shot being taken as it beeps. He is now suspicious of something being on his phone."

Despite being repeatedly told that people were using his product to spy on others without their consent, Fleming helped them with tech support.

A government investigator even opened up an affiliate marketing account for pcTattletale, and Fleming reached out to offer ready-made banner ads with text like “pcTattletale Cheating Husband? #1 catch a cheater spy tracker" and "pcTattletale Husband Cheating? Best Catch a Cheater Spy App."

Fleming noted in an email that pcTattletale was more successful when marketed at women, because "There are a lot more women wanting to catch their man then [sic] the other way around." Financial records showed that Fleming was selling around 1,200 pcTattletale subscriptions a year at anywhere from $99 to $300.

Based on all this, the government obtained a search warrant in late 2022 and raided the Bruce Township home where Fleming lived.

In 2024, TechCrunch reported that pcTattletale was hacked and much of its data was leaked. Apparently, hackers had gained access to the company's private keys for the Amazon Web Services account where most of the video data created by the app was stored. Fleming claimed at the time that his company was “out of business and completely done” after the breach.

The feds eventually charged Fleming with selling a product while "knowing or having reason to know" that the software was "primarily useful for the purpose of the surreptitious interception of wire, oral, or electronic communications." This week in California, Fleming pled guilty to a single count and was released on his own recognizance while awaiting sentencing.

One piece of stalkerware is off the market; unfortunately, many others remain, and their owners and operators are often harder to find.

Read full article

Comments