Ivy League college students are, by definition, intelligent. They don't need to use generative AI to cheat on exams; they could just learn the material. But they also tend to be competitive, ambitious, and overscheduled, so AI can look like an easy shortcut that makes more time in their lives for things that can't be done by a chatbot. When the pressure is on, which approach do they choose?

A new scandal at Brown University reveals that huge numbers of these students are likely to cheat.

Record scores

A recent survey of Princeton students found that 29.9 percent admitted to cheating with AI on at least one exam or assignment. But the situation at Brown gives us a better sense of what this kind of cheating looks like in one particular class—and just how much it may be substituting for actual learning. And we know all this because the blind economics professor at the center of it all, Roberto Serrano, is not letting it go.

Read full article

Comments

One of the most horrific cases of allegedly Grok-generated child sex images was shared in a proposed class action lawsuit that was expanded Tuesday. Now, young girls not only accuse X and xAI of building toxic AI "nudify" tools but also of shielding child predators by obstructing police investigations into Grok-generated child sex abuse materials (CSAM).

In March, a girl’s stepfather took his own life after cops discovered that he had used Grok to create 7,000 sexually explicit images using one photo taken when his stepdaughter was 11 years old, the amended complaint alleged.

Grok allowed the man to generate extreme images depicting incest and rape without flagging any harmful behavior, the complaint said. Seemingly, xAI’s child safety system only intervened after the man input a prompt for “gang rape.” That request sent a CyberTip to the National Center for Missing and Exploited Children (NCMEC), which alerted law enforcement to the AI CSAM.

Read full article

Comments

Kalshi lost an attempt to override New York's state gambling laws yesterday, with a federal judge rejecting the prediction market operator's request to prevent enforcement of the rules.

Kalshi is appealing the decision to a higher court. This is one of numerous cases in which judges must decide whether state laws are preempted by federal regulation of prediction markets.

New York Governor Kathy Hochul and Attorney General Letitia James issued a joint statement on the ruling today. “New York’s gambling laws are designed to protect consumers," they said. "Kalshi tried to ignore them. Yesterday, they lost in court. We will continue to hold all gambling platforms accountable to the law—and that includes prediction markets.”

Read full article

Comments

A Linux vulnerability that allows untrusted virtual machines to gain root access to host machines is one of two high-severity flaws to surface this week in the open source operating system.

The vulnerability resides in KVM, which is, in essence, a virtual machine app included in the kernel of many Linux distributions. The vulnerability, tracked as CVE-2026-53359, allows guest virtual machines—such as those used in cloud platforms to isolate one user’s instance from the host OS and other user instances—to break out of that container.

Januscape: A threat to cloud platforms

The vulnerability affects KVM running on both AMD and Intel processors. It exploits bugs residing in the KVM guest-side, the portion of the VM that consists of only resources like the OS or drivers present in the guest VM, rather than resources present on the host machine. The threat went unnoticed in the Linux kernel for 16 years.

Read full article

Comments

Last week, thousands of SamKnows routers were bricked after a government program ran its course.

In 2020, as part of a program conducted by the Australian Competition & Consumer Commission (ACCC), the Australian government's chief competition regulator, thousands of volunteers received routers to help test and report on the typical speed and performance of broadband plans in Australia. (More specifically, the Measuring Broadband Australia (MBA) program targeted fixed-line broadband services provided over the NBN, Australia's government-owned wholesale open-access broadband network, as well as services delivered over other access networks.)

According to the final report that the ACCC distributed, the routers are whiteboxes that were “supplied by SamKnows” and that “perform tests to measure internet performance using test servers maintained by SamKnows and hosted in Australia.”

Read full article

Comments

TikTok's For You Page (FYP) is the default home screen for users of the video-sharing platform. It's a personalized, algorithmically driven content feed, but the approach differs from other social media in that TikTok's algorithm relies heavily on implicit signals—such as how long users watch particular videos—as well as explicit signals such as likes or follows. And generally, that algorithm does remarkably well at predicting which videos will interest particular users.

But some users have voiced concerns that TikTok's almighty algorithm doesn't seem to incorporate negative feedback very well. Even when they don't watch a suggested video or click the "not interested" feature, they keep seeing those videos on their FYP. Northeastern University computer scientists put those suspicions to the test. According to their recent paper, the engagement signals do have an effect, but only temporarily. Then the algorithm gradually relapses unless a user consistently gives the same feedback over and over again.

The research group specializes in "algorithm audits," co-author Piotr Sapiezynski told Ars, to better understand online platforms: "how they work, how they fail, when they fail, how they harm individuals and societies." In this case, he and his co-authors wanted to take a closer look at user agency after hearing multiple anecdotal reports from TikTok users that their negative feedback—responding to prompts by indicating they aren't interested or want to see less of a certain kind of video—doesn't seem to remove those posts from their FYP. "On the other hand, it's unclear why the platforms would offer it, if it doesn't work," said Sapiezynski.

Read full article

Comments

The US military has lost dozens of Reaper drones collectively worth more than $1 billion while carrying out surveillance and attack missions over Iran. Now the Pentagon is seeking large numbers of cheaper drones that can perform such missions despite the expectation that many will be lost in combat.

In a call for industry pitches, the Defense Innovation Unit’s notice described the US military’s current reliance on drones and crewed aircraft, each costing more than $30 million, as being “unsustainable against adversaries utilizing layered defenses enabled by increasingly low-cost antiaircraft capabilities.” It envisions deploying more “cost-effective” drones to “overwhelm enemy air defenses even while experiencing numerous [drone] losses.”

That is, in practice, what Ukraine’s military has been demonstrating with its long- and mid-range strike campaign against Russian supply lines, oil refineries, and various energy or industrial targets within Russia or occupied Ukraine. The Ukrainian campaign has been overwhelming Russia’s overstretched air defense capabilities by launching hundreds of relatively inexpensive drones and missiles on a daily basis to attack targets far behind the frontlines, while continuing to damage or destroy Russia’s most sophisticated air defense systems.

Read full article

Comments

The proliferation of nuclear power in space got a little more real Tuesday with the launch of a small satellite developed by a Florida-based company specializing in nuclear micro-power technology.

It's a long way from launching a bona fide nuclear reactor, a breakthrough that could help power a permanent Moon base and efficiently drive rockets throughout the Solar System. But you have to start somewhere.

The satellite from Miami-based City Labs is named BOHR, short for Betavoltaic Orbital High-Reliability, and it launched on a SpaceX rideshare mission Tuesday alongside 80 other payloads. SpaceX's Falcon 9 rocket released the BOHR satellite into an orbit between 350 and 400 miles (nearly 600 km) in altitude.

Read full article

Comments

Code generation is emerging as one of the most popular applications for large language models (LLMs), but not all agents are equally good at all development tasks. Google created a benchmark earlier this year to evaluate how LLMs perform in Android app development, and Android Bench is getting a big update today. The leaderboard now includes a raft of new models, and Google has adopted a new framework that should be easier to use. Developers are invited to run their own tests and submit feedback that could shape the future of Android Bench.

While they are popular coding tools, LLMs don't get everything right. Separating the useful outputs from straight-up slop means choosing the right tool. Android Bench aims to demonstrate which AI agents do best on a suite of 100 Android development tasks. After launching Android Bench in March, Google has added metrics like cost and efficiency, as well as open-weight models.

To keep Android Bench relevant, Google is updating the test with eight new models, including all the latest heavy-hitters: Claude Fable 5, Claude Sonnet 5, Claude Opus 4.8, GLM 5.2, Kimi K2.7 Code, MiniMax M3, Qwen 3.7 Plus, and Qwen 3.7 Max.

Read full article

Comments

Two California teens have learned that Waymo's robotaxis can and will enforce their rider rules on misbehaving passengers. It probably seemed like a marvelous jape—ride around in the back of a robotaxi getting drunk and doing drive-bys, shooting stuff with gel beads. And perhaps it was, until that robotaxi and Waymo worked out what was going on. Waymo then stopped the car and alerted the San Mateo Police, who showed up and detained the troublemakers, according to a post on the department's Facebook page.

"After calling us and stopping the car, we were able to safely remove both subjects and determined they were shooting Orbeez from the car as they sipped on afternoon libations while being chauffeured around town in the driverless vehicle," the police wrote.

This isn't the first instance of a Waymo robotaxi snitching on misbehaving passengers. According to Reddit, last year two men in Los Angeles were reported to the police for drinking inside a robotaxi.

Read full article

Comments

One of the central features of plate tectonics is the formation of new crust at mid-ocean ridges. Part of the spreading process that drives continents apart, it was arguably the discovery of these ridges that drove widespread acceptance of plate tectonics as a theory. Thanks to decades of exploration, we now have a good picture of what the crust that forms at the site of spreading looks like. But we still have an incomplete idea of how its features are actually produced.

In other words, we have a good idea of the outcome of the process, but not a detailed picture of the process itself.

That is starting to change. In 2024, a team of French scientists was able to remotely monitor a major event on the border between the Australian and Antarctic plates, only two months after they installed equipment on the ocean floor. Their data shows that most of the spreading occurred in a relatively short time window, and some key events happened without any obvious seismic activity.

Read full article

Comments

US rare earths produced by Washington-backed companies are flowing to Japan and South Korea, as American demand has yet to materialize despite the Trump administration’s push to develop a national supply chain.

Rare earths products produced by MP Materials, Energy Fuels and Phoenix Tailings—which together have won billions of dollars in US government support—are being sold to companies in Asia, where the scale of magnet manufacturing remains larger than the nascent production in the US.

China’s lock on global supplies of rare earths and critical minerals has become a national security concern in the US and other Western nations, since Beijing started restricting access to them. The metals are crucial to 21st-century technology and are used in the manufacturing of everything from weapons guidance systems to electric vehicle batteries.

Read full article

Comments

In a world where self-driving robotaxis glide through major city streets without drivers behind the wheel and delivery drones autonomously fly through the skies to drop off orders at customers’ homes, the idea of general-purpose robots helping humans with various tasks in workplaces or even homes may not seem far-fetched.

But that future hinges on developing increasingly autonomous robots powered by modern artificial intelligence—an ambitious vision that has motivated many researchers to become startup founders while also attracting billions of dollars in investment.

“When I started maybe about 15 years ago, I led a project team that was focused on autonomy, but in that era, the goal of that team was to just get a robot to navigate from point A to point B,” said Matt Malchano, vice president of software at the robotics company Boston Dynamics based in Waltham, Massachusetts. “And now, when we think of autonomy, we think of this huge space of tasks and things that we can imagine a robot doing on its own.”

Read full article

Comments

The rocket company founded by Jeff Bezos, Blue Origin, is raising private capital, the DealBook newsletter reported early Wednesday.

According to the publication, the company is raising $10 billion, leading to a valuation of $130 billion. Coatue Management, a big asset manager, is expected to lead with a $4 billion commitment. Another $4 billion is expected to come from large institutional investors. And Bezos will contribute an additional $2 billion.

Founded in 2000, Blue Origin is seeking to become a global leader in spaceflight, developing a line of super heavy lift rockets, lunar landers, and plans for two megaconstellations. It is seeking to compete in the same areas—launch, telecommunications, data centers from space—as SpaceX.

Read full article

Comments

In the brief history of AI security, the prompt injection has quickly become the top threat. Large language models are inherently unable to distinguish between legitimate instructions provided by users and malicious ones sneaked into emails, source code, and other third-party content the models are processing. This makes it trivial to surreptitiously inject malicious commands that the LLM readily follows.

With no way to enforce this crucial boundary between trusted and untrusted sources, AI engine developers are left to erect elaborate guardrails designed to mitigate the damage rather than solve the root cause.

To date, most prompt injections have fallen into a class known as push, in which each potential victim is targeted. For example, the adversary injects malicious instructions into an individual email or calendar invitation. Because the injection must then be sent (or pushed) to each specific target, the scale of the attack is limited, hampering mass exploits that hit the Internet at large.

Read full article

Comments