One developer tells MIT Technology Review that AI tools weaken the coding instincts he used to have. And beyond that, "It's just not fun sitting there with my work being done for me."

But is AI making coders faster? "After speaking to more than 30 developers, technology executives, analysts, and researchers, MIT Technology Review found that the picture is not as straightforward as it might seem..." For some developers on the front lines, initial enthusiasm is waning as they bump up against the technology's limitations. And as a growing body of research suggests that the claimed productivity gains may be illusory, some are questioning whether the emperor is wearing any clothes.... Data from the developer analytics firm GitClear shows that most engineers are producing roughly 10% more durable code — code that isn't deleted or rewritten within weeks — since 2022, likely thanks to AI. But that gain has come with sharp declines in several measures of code quality. Stack Overflow's survey also found trust and positive sentiment toward AI tools falling significantly for the first time. And most provocatively, a July study by the nonprofit research organization Model Evaluation & Threat Research (METR) showed that while experienced developers believed AI made them 20% faster, objective tests showed they were actually 19% slower...

Developers interviewed by MIT Technology Review generally agree on where AI tools excel: producing "boilerplate code" (reusable chunks of code repeated in multiple places with little modification), writing tests, fixing bugs, and explaining unfamiliar code to new developers. Several noted that AI helps overcome the "blank page problem" by offering an imperfect first stab to get a developer's creative juices flowing. It can also let nontechnical colleagues quickly prototype software features, easing the load on already overworked engineers. These tasks can be tedious, and developers are typically glad to hand them off. But they represent only a small part of an experienced engineer's workload. For the more complex problems where engineers really earn their bread, many developers told MIT Technology Review, the tools face significant hurdles...

The models also just get things wrong. Like all LLMs, coding models are prone to "hallucinating" — it's an issue built into how they work. But because the code they output looks so polished, errors can be difficult to detect, says James Liu, director of software engineering at the advertising technology company Mediaocean. Put all these flaws together, and using these tools can feel a lot like pulling a lever on a one-armed bandit. "Some projects you get a 20x improvement in terms of speed or efficiency," says Liu. "On other things, it just falls flat on its face, and you spend all this time trying to coax it into granting you the wish that you wanted and it's just not going to..." There are also more specific security concerns, she says. Researchers have discovered a worrying class of hallucinations where models reference nonexistent software packages in their code. Attackers can exploit this by creating packages with those names that harbor vulnerabilities, which the model or developer may then unwittingly incorporate into software.
Other key points from the article:

• LLMs can only hold limited amounts of information in context windows, so "they struggle to parse large code bases and are prone to forgetting what they're doing on longer tasks."

• "While an LLM-generated response to a problem may work in isolation, software is made up of hundreds of interconnected modules. If these aren't built with consideration for other parts of the software, it can quickly lead to a tangled, inconsistent code base that's hard for humans to parse and, more important, to maintain."

• "Accumulating technical debt is inevitable in most projects, but AI tools make it much easier for time-pressured engineers to cut corners, says GitClear's Harding. And GitClear's data suggests this is happening at scale..."

• "As models improve, the code they produce is becoming increasingly verbose and complex, says Tariq Shaukat, CEO of Sonar, which makes tools for checking code quality. This is driving down the number of obvious bugs and security vulnerabilities, he says, but at the cost of increasing the number of 'code smells' — harder-to-pinpoint flaws that lead to maintenance problems and technical debt."

Yet the article cites a recent Stanford University study that found employment among software developers aged 22 to 25 dropped nearly 20% between 2022 and 2025, "coinciding with the rise of AI-powered coding tools."

The story is part of MIT Technology Review's new Hype Correction series of articles about AI.

Meta is informing some users that they will soon be restricted in how many link posts they can share each month, unless they pay for its Meta Verified subscription service. As per the notification message: "Starting December 16, certain Facebook profiles without Meta Verified, including yours, will be limited to sharing links in 2 organic posts per month. Subscribe to Meta Verified to share more links on Facebook, plus get a verified badge and additional benefits to help protect your brand."

To be clear, right now this is a limited test, so relatively few Pages are impacted. But understandably, a lot of users are also seeking more information on the change, and whether it could be expanded to all Pages. So, Meta's seeking to boost take-up of Meta Verified, in order to make more money out of its subscription option, which, for business users, costs between $14.99 and $499 per month, depending on which package you choose.

U.S. senators are probing whether Big Tech data centers are driving up local electricity bills by socializing grid upgrade costs onto residents. Some of the tactics they're using include NDAs, shell companies, and lobbying. Ars Technica reports: In letters (PDF) to seven AI firms, Senators Elizabeth Warren (D-Mass.), Chris Van Hollen (D-Md.), and Richard Blumenthal (D-Conn.) cited a study estimating that "electricity prices have increased by as much as 267 percent in the past five years" in "areas located near significant data center activity." Prices increase, senators noted, when utility companies build out extra infrastructure to meet data centers' energy demands -- which can amount to one customer suddenly consuming as much power as an entire city. They also increase when demand for local power outweighs supply. In some cases, residents are blindsided by higher bills, not even realizing a data center project was approved, because tech companies seem intent on dodging backlash and frequently do not allow terms of deals to be publicly disclosed.

AI firms "ask public officials to sign non-disclosure agreements (NDAs) preventing them from sharing information with their constituents, operate through what appear to be shell companies to mask the real owner of the data center, and require that landowners sign NDAs as part of the land sale while telling them only that a 'Fortune 100 company' is planning an 'industrial development' seemingly in an attempt to hide the very existence of the data center," senators wrote. States like Virginia with the highest concentration of data centers could see average electricity prices increase by another 25 percent by 2030, senators noted. But price increases aren't limited to the states allegedly striking shady deals with tech companies and greenlighting data center projects, they said. "Interconnected and interstate power grids can lead to a data center built in one state raising costs for residents of a neighboring state," senators reported.

Under fire for supposedly only pretending to care about keeping neighbors' costs low were Amazon, Google, Meta, Microsoft, Equinix, Digital Realty, and CoreWeave. Senators accused firms of paying "lip service," claiming that they would do everything in their power to avoid increasing residential electricity costs, while actively lobbying to pass billions in costs on to their neighbors. [...] Particularly problematic, senators emphasized, were reports that tech firms were getting discounts on energy costs as utility companies competed for their business, while prices went up for their neighbors.

Intel has quietly stopped maintaining its open-source user-space driver stack for Gaudi accelerators. Phoronix reports: It turns out earlier this year Intel archived the SynapseAI Core open-source code and is no longer maintained by Intel. The open-source Synapse AI Core GitHub repository was archived in February and README updated with: "This project will no longer be maintained by Intel. Intel has ceased development and contributions including, but not limited to, maintenance, bug fixes, new releases, or updates, to this project. Intel no longer accepts patches to this project. If you have an ongoing need to use this project, are interested in independently developing it, or would like to maintain patches for the open source software community, please create your own fork of this project."

An anonymous reader quotes a report from BleepingComputer: Audio streaming platform SoundCloud has confirmed that outages and VPN connection issues over the past few days were caused by a security breach in which threat actors stole a database containing user information. The disclosure follows widespread reports over the past four days from users who were unable to access SoundCloud when connecting via VPN, with attempts resulting in the site displaying 403 "forbidden" errors.

In a statement shared with BleepingComputer, SoundCloud said it recently detected unauthorized activity involving an ancillary service dashboard and activated its incident response procedures. SoundCloud acknowledged that a threat actor accessed some of its data but said the exposure was limited in scope. [...] BleepingComputer has learned that the breach affects 20% of SoundCloud's users, which, based on publicly reported user figures, could impact roughly 28 million accounts. The company said it is confident that all unauthorized access to SoundCloud systems has been blocked and that there is no ongoing risk to the platform. "We understand that a purported threat actor group accessed certain limited data that we hold," SoundCloud told BleepingComputer. "We have completed an investigation into the data that was impacted, and no sensitive data (such as financial or password data) has been accessed. The data involved consisted only of email addresses and information already visible on public SoundCloud profiles."

A new study warns that glaciers in the European Alps will hit their peak extinction rate within eight years, with global glacier loss accelerating toward thousands per year unless emissions are rapidly cut. "Glaciers in the western US and Canada are forecast to reach their peak year of loss less than a decade later, with more than 800 disappearing each year by then," adds the Guardian. From the report: About 200,000 glaciers remain worldwide, with about 750 disappearing each year. However, the research indicates this pace will accelerate rapidly as emissions from burning fossil fuels continue to be released into the atmosphere. Current climate action plans from governments are forecast to push global temperatures to about 2.7C above preindustrial levels, supercharging extreme weather. Under this scenario, glacier losses would peak at about 3,000 a year in 2040 and plateau at that rate until 2060. By the end of the century, 80% of today's glaciers will have gone. By contrast, rapid cuts to carbon emissions to keep global temperature rise to 1.5C would cap annual losses at about 2,000 a year in 2040, after which the rate would decline. [...]

The new study, published in Nature Climate Change, analyzed more than 200,000 glaciers from a database of outlines derived from satellite images. The researchers used three global glacier models to assess their fate under different heating scenarios. Regions with the smallest and fastest-melting glaciers were found to be the most vulnerable. The study estimates the 3,200 glaciers in central Europe would shrink by 87% by 2100 -- even if global temperature rise is limited to 1.5C, rising to 97% under 2.7C of heating.

In the western US and Canada, including Alaska, about 70% of today's 45,000 glaciers are projected to vanish under 1.5C of heating, and more than 90% under 2.7C. The Caucasus and southern Andes are also expected to face devastating losses. Larger glaciers take longer to melt, with those in Greenland reaching their peak extinction rate in about 2063 -- losing 40% by 2100 under 1.5C of heating and 59% under 2.7C. However, the melting is forecast to continue beyond 2100. The researchers said the peak loss dates represent more than a numerical milestone. "They mark turning points with profound implications for ecosystems, water resources and cultural heritage," they wrote. "[It is] a human story of vanishing landscapes, fading traditions and disrupted daily routines."

The R programming language "is sometimes frowned upon by 'traditional' software engineers," says the CEO of software quality services vendor Tiobe, "due to its unconventional syntax and limited scalability for large production systems." But he says it "continues to thrive at universities and in research-driven industries, and "for domain experts, it remains a powerful and elegant tool."

Yet it's now gaining more popularity as statistics and large-scale data visualization become important (a trend he also sees reflected in the rise of Wolfram/Mathematica). That's according to December's edition of his TIOBE Index, which attempts to rank the popularity of programming languages based on search-engine results for courses, third-party vendors, and skilled engineers. InfoWorld explains: In the December 2025 index, published December 7, R ranks 10th with a 1.96% rating. R has cracked the Tiobe index's top 10 before, such as in April 2020 and July 2020, but not in recent years. The rival Pypl Popularity of Programming Language Index, meanwhile, has R ranked fifth this month with a 5.84% share. "Programming language R is known for fitting statisticians and data scientists like a glove," said Paul Jansen, CEO of software quality services vendor Tiobe, in a bulletin accompanying the December index...

Although data science rival Python has eclipsed R in terms of general adoption, Jansen said R has carved out a solid and enduring niche, excelling at rapid experimentation, statistical modeling, and exploratory data analysis. "We have seen many Tiobe index top 10 entrants rising and falling," Jansen wrote. "It will be interesting to see whether R can maintain its current position."
"Python remains ahead at 23.64%," notes TechRepublic, "while the familiar chase group behind it holds steady for the moment. The real movement comes deeper in the list, where SQL edges upward, R rises to the top 10, and Delphi/Object Pascal slips away... SQLclimbs from tenth to eighth at 2.10%, adding a small +0.11% that's enough to move it upward in a tightly packed section of the table. Perl holds ninth at 1.97%, strengthened by a +1.33% gain that extends its late-year resurgence."

It's interesting to see how TIOBE's ranking compare with PYPL's (which ranks languages based solely on how often language tutorials are searched on Google):

TIOBE	PYPL
Python	Python
C	C/C++
C++	Objective-C
Java	Java
C#	R
JavaScript	JavaScript
Visual Basic	Swift
SQL	C#
Perl	PHP
R	Rust

Despite their different methodologies, both lists put Python at #1, Java at #5, and JavaScript at #7.

New submitter emangwiro shares a report: The HDMI Forum, responsible for the HDMI specification, continues to stonewall open source. Valve's Steam Machine theoretically supports HDMI 2.1, but the mini-PC is software-limited to HDMI 2.0. As a result, more than 60 frames per second at 4K resolution are only possible with limitations. In a statement to Ars Technica, a Valve spokesperson confirmed that HDMI 2.1 support is "still a work-in-progress on the software side." "We've been working on trying to unblock things there."

The Steam Machine uses an AMD Ryzen APU with a Radeon graphics unit. Valve strictly adheres to open-source drivers, but the HDMI Forum is unwilling to disclose the 2.1 specification. According to Valve, they have validated the HDMI 2.1 hardware under Windows to ensure basic functionality.

An anonymous reader quotes a report from Reuters: Meta's proposal to use less personal data for targeted advertising in its pay-or-consent model that will be rolled out next month won the approval of EU antitrust regulators on Monday, signaling the company will not face daily fines after all. [...] The U.S. tech giant has been locked in discussions with the European Commission after getting hit with a $233 million fine in April for breaching the Digital Markets Act aimed at reining in the power of Big Tech. The violation covered Facebook and Instagram in the period from November 2023 to November 2024, after which Meta tweaked its pay-or-consent model to use less personal data for targeted advertising.

The EU executive has been examining the changes to see if they comply with the DMA, with Meta risking daily fines of as much as 5% of its average daily worldwide turnover if found to be still in breach of the law. The tweaks are in wording, design and transparency to remind users of the two options. Meta did not plan on any substantial changes to its November proposal despite the risk of EU fines, people with direct knowledge of the matter had told Reuters. The Commission, which acts as the EU competition enforcer, acknowledged Meta's November proposal, saying that it will monitor the new ad model and seek feedback, with no more talk of periodic fines. "Meta will give users the effective choice between consenting to share all their data and seeing fully personalized advertising, and opting to share less personal data for an experience with more limited personalized advertising," the Commission said in a statement.

UPDATE (12/7): The New York Times clarifies today that the damage at Chernobyl hasn't led to a rise in radiation levels: "If there was to be some event inside the shelter that would release radioactive materials into the space inside the New Safe Confinement, because this facility is no longer sealed to the outside environment, there's the potential for radiation to come out," said Shaun Burnie, a senior nuclear specialist at Greenpeace who has monitored nuclear power plants in Ukraine since 2022 and last visited Chernobyl on October 31. "I have to say I don't think that's a particularly serious issue at the moment, because they're not actively decommissioning the actual sarcophagus."

The I.A.E.A. also said there was no permanent damage to the shield's load-bearing structures or monitoring systems. A spokesman for the agency, Fredrik Dahl, said in a text message on Sunday that radiation levels were similar to what they were before the drone hit.
But "A structure designed to prevent radioactive leakage at the defunct Chernobyl nuclear plant in Ukraine is no longer operational," Politico reported Saturday, "after Russian drones targeted it earlier this year, the U.N.'s nuclear watchdog has found." [T]he large steel structure "lost its primary safety functions, including the confinement capability" when its outer cladding was set ablaze after being struck by Russian drones, according to a new report by the International Atomic Energy Agency. Beyond that, there was "no permanent damage to its load-bearing structures or monitoring systems," it said. "Limited temporary repairs have been carried out on the roof, but timely and comprehensive restoration remains essential to prevent further degradation and ensure long-term nuclear safety," IAEA Director General Rafael Mariano Grossi said in astatement.
The Guardian has pictures of the protective shield — incuding the damage from the drone strike. The shield is the world's largest movable land structure, reports CNN: The IAEA, which has a permanent presence at the site, will "continue to do everything it can to support efforts to fully restore nuclear safety and security," Grossi said.... Built in 2010 and completed in 2019, it was designed to last 100 years and has played a crucial role in securing the site.

The project cost €2.1 billion and was funded by contributions from more than 45 donor countries and organizations through the Chernobyl Shelter Fund, according to the European Bank for Reconstruction and Development, which in 2019 hailed the venture as "the largest international collaboration ever in the field of nuclear safety."

"Homebrew, the package manager for macOS and Linux, just got a handy new feature in the latest v5.0.4 update," reports How-To Geek.

Brewfile install scripts "are now more like a one-stop shop for installing software, as Flatpaks are now supported alongside Brew packages, Mac App Store Apps, and other packages." For those times when you need to install many software packages at once, like when setting up a new PC or virtual machine, you can create a Brewfile with a list of packages and run it with the 'brew bundle' command. However, the Brewfile isn't limited to just Homebrew packages. You can also use it to install Mac App Store apps, graphical apps through Casks, Visual Studio Code extensions, and Go language packages. Starting with this week's Homebrew v5.0.4 release, Flatpaks are now supported in Brewfiles as well...

This turns Homebrew into a fantastic setup tool for macOS, Linux, and Windows Subsystem for Linux (WSL) environments. You can have one script with all your preferred software, and use 'if' statements with platform variables and existing file checks for added portability.

India is weighing a proposal to mandate always-on satellite tracking in smartphones for precise government surveillance -- an idea strongly opposed by Apple, Google, Samsung, and industry groups. Reuters reports: For years, the [Prime Minister Narendra Modi's] administration has been concerned its agencies do not get precise locations when legal requests are made to telecom firms during investigations. Under the current system, the firms are limited to using cellular tower data that can only provide an estimated area location, which can be off by several meters.

The Cellular Operators Association of India (COAI), which represents Reliance's Jio and Bharti Airtel, has proposed that precise user locations should only be provided if the government orders smartphone makers to activate A-GPS technology -- which uses satellite signals and cellular data -- according to a June internal federal IT ministry email. That would require location services to always be activated in smartphones with no option for users to disable them. Apple, Samsung, and Alphabet's Google have told New Delhi that should not be mandated, said three of the sources who have direct knowledge of the deliberations.

A measure to track device-level location has no precedent anywhere else in the world, lobbying group India Cellular & Electronics Association (ICEA), which represents both Apple and Google, wrote in a confidential July letter to the government, which was viewed by Reuters. "The A-GPS network service ... (is) not deployed or supported for location surveillance," said the letter, which added that the measure "would be a regulatory overreach." Earlier this week, Modi's government was forced to rescind an order requiring smartphone makers to preload a state-run cyber safety app on all devices after public backlash and privacy concerns.

Airbus said Monday that the vast majority of around 6,000 A320-family jets affected by an emergency software recall have now been modified, leaving fewer than 100 aircraft still requiring work after a frantic weekend of repairs prompted by the discovery of a vulnerability to solar flares. The unprecedented recall -- described as the broadest emergency action in the company's history -- came after a mid-air incident on a JetBlue A320 revealed a possible link between a drop in altitude and a space-related computer bug.

The fix involved reverting to an earlier version of software that controls nose angle, uploaded via cable from a portable device called a data loader. Some older A320 jets will need entirely new computers rather than a simple software reset, raising questions about how long those aircraft will remain grounded amid global chip shortages.

Reuters separately reported on Monday that Airbus had discovered an industrial quality issue affecting metal panels of a "limited" number of A320-family aircraft. The company told the publication that it had "identified" and "contained" the source of the issue and that "all newly produced panels conform to all requirements."

A gauge of risk on Oracle debt "reached a three-year high in November," reports Bloomberg.

"And things are only going to get worse in 2026 unless the database giant is able to assuage investor anxiety about a massive artificial intelligence spending spree, according to Morgan Stanley." A funding gap, swelling balance sheet and obsolescence risk are just some of the hazards Oracle is facing, according to Lindsay Tyler and David Hamburger, credit analysts at the brokerage.

The cost of insuring Oracle's debt against default over the next five years rose to 1.25 percentage point a year on Tuesday, according to ICE Data Services. The price on the five-year credit default swaps is at risk of toppling a record set in 2008 as concerns over the company's borrowing binge to finance its AI ambitions continue to spur heavy hedging by banks and investors, they warned in a note Wednesday. The CDS could break through 1.5 percentage point in the near term and could approach 2 percentage points if communication around its financing strategy remains limited as the new year progresses, the analysts wrote. Oracle CDS hit a record 1.98 percentage point in 2008, ICE Data Services shows...

"Over the past two months, it has become more apparent that reported construction loans in the works, for sites where Oracle is the future tenant, may be an even greater driver of hedging of late and going forward," wrote the analysts... Concerns have also started to weigh on Oracle's stock, which the analysts said may incentivize management to outline a financing plan on the upcoming earnings call...
Thanks to Slashdot reader Bruce66423 for sharing the article.

"OpenAI is now internally testing 'ads' inside ChatGPT," reports BleepingComputer: Up until now, the ChatGPT experience has been completely free. While there are premium plans and models, you don't see GPT sell you products or show ads. On the other hand, Google Search has ads that influence your buying behaviour. OpenAI is planning to replicate a similar experience.

As spotted [by software engineer Tibor Blaho] on X.com,ChatGPT Android app 1.2025.329 beta includes new references to an "ads feature" with "bazaar content", "search ad" and "search ads carousel."

This move could disrupt the web economy, as what most people don't understand is that GPT likely knows more about users than Google. For example, OpenAI could create personalised ads on ChatGPT that promote products that you really want to buy... The leak suggests that ads will initially be limited to the search experience only, but this may change in the future.

Poland has launched a new antitrust investigation into Apple's App Tracking Transparency rules, questioning whether Apple misled users about privacy while giving its own apps a competitive advantage over third-party developers. AppleInsider reports: On November 25, Poland's UOKiK has started another investigation into App Tracking Transparency, and whether Apple had restricted competition in mobile advertising. Reuters reports that, to the anti-monopoly regulator, ATT may have limited advertisers' ability to collect user data for advertising purposes while simultaneously favoring Apple's ad program. On November 25, Poland's UOKiK has started another investigation into App Tracking Transparency, and whether Apple had restricted competition in mobile advertising. Reuters reports that, to the anti-monopoly regulator, ATT may have limited advertisers' ability to collect user data for advertising purposes while simultaneously favoring Apple's ad program.

This is not the first time that Poland has looked into ATT rules. In December 2021, the regulator held a similar probe following criticism from advertisers. It's not clear what that complaint determined, or if it is still ongoing. Regardless, in the new complaint, the logic is that Apple had a competitive advantage since its own apps were not subject to ATT rules, but third-party apps did have to deal with ATT. Since Apple didn't visibly ask for consent for its first-party apps in the same way, there is a presumption that Apple's rules only applied to other companies.

This is despite Apple's repeated insistence that it doesn't use the same kinds of collected data in its own apps and services for marketing purposes, as well as its stance on privacy in general. In short, Apple apps don't use the data, so it doesn't pop up a dialog box asking the user if the app can use the data. There is also the argument that, in setting up an account with Apple, users are providing blanket consent to the company. Implementing ATT on its own apps would therefore be a waste of time, since that consent was already granted. Apple said that it will work with the regulator on the matter, but warned that it could force them to withdraw the feature "to the detriment of European consumers."

The abstract of a paper on NBER: This paper examines a large-scale randomized evaluation of the One Laptop Per Child (OLPC) program in 531 Peruvian rural primary schools. We use administrative data on academic performance and grade progression over 10 years to estimate the long-run effects of increased computer access on (i) school performance over time and (ii) students' educational trajectories. Following schools over time, we find no significant effects on academic performance but some evidence of negative effects on grade progression. Following students over time, we find no significant effects on primary and secondary completion, academic performance in secondary school, or university enrollment. Survey data indicate that computer access significantly improved students' computer skills but not their cognitive skills; treated teachers received some training but did not improve their digital skills and showed limited use of technology in classrooms, suggesting the need for additional pedagogical support.

Google's long-rumored plan to merge ChromeOS and Android into a single desktop operating system now has a name: Aluminium OS, AndroidAuthority reports, citing a job listing.

The job listing explicitly tasks applicants with "working on a new Aluminium, Android-based, operating system." The job listing confirms Google intends to eventually replace ChromeOS entirely, though the two platforms will coexist during a transition period. Aluminium OS won't be limited to budget hardware -- the listing references "AL Entry," "AL Mass Premium," and "AL Premium" tiers across laptops, detachables, tablets, and mini-PCs.

An anonymous reader quotes a report from the Associated Press: They're cute, even cuddly, and promise learning and companionship -- but artificial intelligence toys are not safe for kids, according to children's and consumer advocacy groups urging parents not to buy them during the holiday season. These toys, marketed to kids as young as 2 years old, are generally powered by AI models that have already been shown to harm children and teenagers, such as OpenAI's ChatGPT, according to an advisory published Thursday by the children's advocacy group Fairplay and signed by more than 150 organizations and individual experts such as child psychiatrists and educators.

"The serious harms that AI chatbots have inflicted on children are well-documented, including fostering obsessive use, having explicit sexual conversations, and encouraging unsafe behaviors, violence against others, and self-harm," Fairplay said. AI toys, made by companies including Curio Interactive and Keyi Technologies, are often marketed as educational, but Fairplay says they can displace important creative and learning activities. They promise friendship but disrupt children's relationships and resilience, the group said. "What's different about young children is that their brains are being wired for the first time and developmentally it is natural for them to be trustful, for them to seek relationships with kind and friendly characters," said Rachel Franz, director of Fairplay's Young Children Thrive Offline Program. Because of this, she added, the trust young children are placing in these toys can exacerbate the types of harms older children are already experiencing with AI chatbots.

A separate report Thursday by Common Sense Media and psychiatrists at Stanford University's medical school warned teenagers against using popular AI chatbots as therapists. Fairplay, a 25-year-old organization formerly known as the Campaign for a Commercial-Free Childhood, has been warning about AI toys for years. They just weren't as advanced as they are today. A decade ago, during an emerging fad of internet-connected toys and AI speech recognition, the group helped lead a backlash against Mattel's talking Hello Barbie doll that it said was recording and analyzing children's conversations. This time, though AI toys are mostly sold online and more popular in Asia than elsewhere, Franz said some have started to appear on store shelves in the U.S. and more could be on the way. "Everything has been released with no regulation and no research, so it gives us extra pause when all of a sudden we see more and more manufacturers, including Mattel, who recently partnered with OpenAI, potentially putting out these products," Franz said. Last week, consumer advocates at U.S. PIRG called out the trend of buying AI toys in its annual "Trouble in Toyland" report. This year, the organization tested four toys that use AI chatbots. "We found some of these toys will talk in-depth about sexually explicit topics, will offer advice on where a child can find matches or knives, act dismayed when you say you have to leave, and have limited or no parental controls," the report said.

Boston Dynamics' Spot robot is now deployed by more than 60 bomb squads and SWAT teams across the US and Canada. The 75-pound four-legged machine starts at around $100,000 and has been used in armed standoffs, hostage rescues and hazardous materials incidents since its commercial debut five years ago. The Massachusetts State Police operates two Spot units purchased in 2020 and 2022. Each cost about $250,000 including add-ons funded through state grants. Last year one of the robots helped corner a suspect who had taken his mother hostage at knifepoint in Hyannis. Houston operates three units and Las Vegas has one.

ICE recently spent around $78,000 on a similar robot from Canadian manufacturer Icor Technology that can also deploy smoke bombs. Civil liberties groups have raised concerns about normalizing militarized policing. The NYPD suspended its limited Spot program in 2021 after public backlash over cost and surveillance concerns before later reinstating it and purchasing two units. The Electronic Frontier Foundation says there should be state and federal laws providing guidance on appropriate use of such technology. About 2,000 Spot units now operate globally.