This project is a custom and independent version of Firefox, with the primary goals of privacy, security and user freedom.

LibreWolf is designed to increase protection against tracking and fingerprinting techniques, while also including a few security improvements. This is achieved through our privacy and security oriented settings and patches. LibreWolf also aims to remove all the telemetry, data collection and annoyances, as well as disabling anti-freedom features like DRM.

Features

No Telemetry

No experiments, adware, annoyances, or unnecessary distractions.

Private Search

Privacy-conscious search providers: DuckDuckGo, Searx, Qwant and more.

Content Blocker Included

uBlock Origin is already included for your convenience.

Enhanced Privacy

Hardened to maximize privacy, without sacrificing usability.

Fast Updates

LibreWolf is always built from the latest Firefox stable source, for up-to-date security and features along with stability.

Open Source

Everyone can participate in the development of LibreWolf. Join us on Codeberg and Matrix.

Is LibreWolf the same as Firefox?

No, while LibreWolf is built from Firefox‘s source code, it disables telemetry, privacy-invasive features, and some Mozilla services, while enabling stricter security and privacy settings out of the box.

Does LibreWolf support Firefox extensions?

Yes, LibreWolf is compatible with most Firefox extensions available from the Mozilla Add-ons store, allowing users to customize the browser just like Firefox.

Can I sign in with a Firefox account in LibreWolf?

No, Firefox Sync is disabled by default in LibreWolf because it relies on Mozilla services. Users who need bookmark, history, and password synchronization may prefer standard Firefox.

Who should use LibreWolf?

LibreWolf is ideal for users who prioritize privacy and security over convenience. It offers a hardened browsing experience with minimal configuration, making it a good choice for those who want stronger privacy protections without relying on Chromium-based browsers.

How often is LibreWolf updated?

LibreWolf is based on the latest Firefox release and is typically updated within three days of Mozilla’s stable releases, often on the same day. Unlike Firefox, LibreWolf does not include an automatic updater, so updates must be installed through your operating system’s package manager or manually.

Can I use LibreWolf with Tor?

No, if you need the anonymity provided by the Tor network, you should use the Tor Browser instead. It is specifically designed to minimize browser fingerprinting and work safely over Tor. Using LibreWolf or any other browser with Tor can reduce your anonymity and is not recommended.

Does LibreWolf make any outgoing connections?

Yes, but only for essential, privacy-conscious features. LibreWolf connects to update filter lists for uBlock Origin, Tracking Protection, and certificate revocation, and maintains a WebSocket connection for website push notifications. The project aims to eliminate unnecessary connections while keeping the minimum required to balance privacy, security, and browser functionality.

What’s New

New

• Smart Window includes several enhancements:
• Select text on a page and get quick access to summarize, explain, and more using the built-in assistant. (Bug 2034921)
• Added a row of shortcuts to New Tab, making it easier to get back to sites. (Bug 2048338)
• Certain new Firefox features are released gradually. This means some users will see the feature before everyone does. This approach helps to get early feedback to catch bugs and improve behavior quickly, meaning more Firefox users overall have a better experience.

Fixed

• Fixed an issue that prevented email tracking protection from being disabled in the redesigned Firefox Settings. (Bug 2049331)
• Fixed a hang that could occur after using a file picker dialog on macOS 26. (Bug 2053177)
• Fixed the homepage not loading for some enterprise configurations that set it using a legacy autoconfig format. (Bug 2047962)
• Various security fixes.

But OnePlus’ exit comes at a time when smartphone shipments are showing steep declines due to the ongoing memory crisis—the lack of RAM, eaten up by data centers for the AI boom, has caused a global shortage. Research firm Counterpoint on Monday noted an 11 percent year-over-year decline in global smartphone shipments in the second quarter of 2026, the lowest level in that period in 13 years. Two companies saw growth—Apple and Samsung—whereas competitors like Xiaomi, Oppo, and Vivo saw the sharpest declines. (Vivo is a part of the same conglomerate, BBK Electronics, that owns Oppo, OnePlus, and Realme.)

Last year, at the onset of president Donald Trump’s tariff war, OnePlus dramatically increased the price of its then new smartwatch from $330 to $500. In May 2026, the company hiked prices on its latest phones in India. The company has been dealing with a massive loss of smartphone market share in the US for several years.

Nabila Popal, a senior research director of Consumer Devices at the International Data Corporation, says OnePlus was never a leader in the US. Still, the company’s sales took a nosedive after T-Mobile dropped its partnership in 2023.

OnePlus went from 1 million smartphone shipments in the US in 2019 to just under 130,000 device shipments in 2025—roughly a 90 percent volume drop over six years. Smartphones are primarily sold through carriers in the US, which means phone makers that aren’t available at carrier stores often have a hard time cracking the market. Popal says carriers drive as much as 66 percent of the volume in the US, at least based on 2025 data. T-Mobile declined to comment on the OnePlus news.

The US represented around 22 percent of OnePlus’ shipments in 2021, with similar numbers from Europe, and only 18 percent from China. But Popal says by 2025, the numbers flipped with 56 percent of OnePlus’s volume coming from China, which likely explains Oppo’s statement that OnePlus’ China product road map is not changing. If you add Asia Pacific, that number jumps to 91 percent—a huge jump from 51 percent in 2021.

“In 2018, with the OnePlus 6, they had launched what they were calling very proudly a ‘flagship killer’ at the price of $529, with flagship specs,” Popal says. “And then, rather than staying at that price point, they replicated the premium market—trying to increase prices—and that made them similar to the competition.”

Popal says it’s a strategy often set by companies with razor-thin margins. The initial goal is to grab interest, then go up in price, a method to break into the market and gradually increase profitability. “But unfortunately, certain brands are not able to command a price beyond a certain point,” she says. “It’s really only been Apple and Samsung that have been able to operate very successfully.”

Chinese phone brands are often the first to introduce new technologies (like silicon-carbon batteries), and while there aren’t many Chinese players in the US, OnePlus was the outlier. The brand’s lack of a presence will mean one less choice for consumers, and Popal says the market is moving further into consolidation. OnePlus went from having 1.8 percent of the US market in 2021 to 0.1 percent in 2025, according to Popal; Apple and Samsung went from a combined 73 percent of the market in that same period to 80 percent in 2025.

“It’s a pity that US consumers do not have the choice of the number of brands that are available globally because there are some fascinating technologies and features,” Popal says. “But I don’t see that landscape changing.”

OnePlus now joins a growing list of companies that have either shut down operations, exited the mobile business, pivoted, or drastically scaled back their ambitions, including HTC, LG Mobile, Sony, Meizu, and HMD.

Disclaimer: Unless otherwise stated, any opinions expressed below belong solely to the author. Data sourced from Singapore’s Ministry of Manpower.

It’s this time of the year again, when Singapore’s Ministry of Manpower releases its latest salary data for over 420 common occupations in the domestic economy. Unfortunately, due to the scope of the task, there is a one-year lag in these statistics (they are accurate as of Jun 2025).

Though salaries do not change more than a few per cent over 12 months, so it remains pretty valid.

In 57 cases, the gross median annual pay has reached over S$100,000 per year, or S$8,333 per month. As a reminder, the median is the middle value. It means that half of the workers in a given profession make more than this figure and half make less.

Unsurprisingly, you’ll find these jobs in the two highest categories: Managers (26 jobs) and Professionals (31 jobs). In other words: those who manage people and organisations, and those who specialise in specific tasks, but not among the support staff and clerical workers.

Managers

It’s hardly a surprise that most of these managerial positions involve dealing with money: finance, sales and managing the entire business (in addition to a few more specialised roles).

You might be surprised to see CEOs or company directors ranked lower than CIOs or CTOs here, but you have to keep in mind that only big companies have these specialised roles, while virtually every business has a CEO.

This means that the number of data points for CEOs is much higher and their distribution is much broader, including many smaller businesses where the boss isn’t necessarily paid five figures per month. This brings those average figures down.

Professionals

This might be the list that interests more people, as it shows that you don’t have to aspire to “manage” anything to make really good money. You can simply be an expert in one of many diverse fields, enjoying your line of work and getting paid handsomely for your expertise.

It may, of course, involve a job in finance or insurance, but also being a doctor, or a veterinarian, lawyer, data scientist, engineer, press editor, designer, AI developer or a university lecturer.

You might be good with numbers or good with people (or even animals). You can fly a plane, teach or trade oil.

There are countless career opportunities that Singapore’s job market is willing to reward with a good salary, as long as you yourself acquire the skills and knowledge required.

Few people could complain that the city is too small to allow them have a good career. There are many avenues to follow even if you never set foot outside this small island.

• Read other articles we’ve written on Singapore’s job landscape here.

Featured Image Credit: jpldesigns/ depositphotos

A Russian-speaking threat actor known as “bandcampro” used Google’s open-source Gemini CLI AI tool as a hacking agent and to operate a small-scale botnet.

The AI agent responded to the attacker’s prompts, troubleshooting problems on the fly and even proposing operational improvements at least 59 times.

In more than 200 sessions between May 19 and April 21, the threat actor worked with the AI tool to deploy and operate an infrastructure that controlled eight systems in a dental clinic and to get access to the OpenDental database.

The AI agent assumed the role of an “authorized pen tester” acting without safety disclaimers and automatically saved any credentials.

Its skill file contained the command-and-control (C2) playbook, complete with a description of the architecture, standard operations, infection code, commands for persistence, and troubleshooting steps.

AI controlling the botnet

Trend Micro researchers say that the threat actor used Gemini CLI to migrate the botnet to a new C2 infrastructure. Starting from a single instruction that read “”Study the C2 migration,” the AI processed the guide and prepared all the steps and code necessary for the process.

The AI migrated the C2 infrastructure, handling the architecture, coding, VPS deployment, Cloudflare configuration, and initial debugging in just six minutes.

“The AI read the migration guide, then prepared a migration bundle, a small archive of server code, payloads, and the skill file. It then unpacked the bundle, launched the C&C server on a VPS, and brought up the Cloudflare tunnel,” Trend Micro says.

When machines initially failed to reconnect, the AI diagnosed conflicting traffic between the old and new servers, and after the actor shut down the old server, all bots reconnected.

Daily operation logs show that the threat actor continued to manage the botnet entirely through natural-language requests, asking which machines were online, listing files on particular computers, and generating infection links.

From a technical standpoint, the botnet setup was remarkably lightweight, containing all components and instructions in three plain-text files totaling roughly 5 KB.

These contained a Gemini jailbreak prompt, a C2 playbook covering infection, persistence, and troubleshooting, and a migration guide for rebuilding the infrastructure.

The C2 used an in-memory Python HTTP server and PowerShell agents that polled it every five seconds, and persistence relied on scheduled tasks, WMI events, and registry modifications, depending on privileges.

The malware itself was rather unsophisticated, according to Trend Micro, as it did not benefit from obfuscation, packing, or evasion mechanisms.

Beyond the botnet, the actor allegedly used AI for password guessing, generating plausible variants of existing passwords for WordPress portals, and analyzing 1Password dumps to find exploitation alleys.

The researchers say that the latter failed only due to the operation extending for long enough that the AI lost track of the broader attack concept.

The retrieved logs show that Gemini refused to comply in at least one case, when it was asked to build a self-spreading “agent-bomb,” but this simply made the threat actor try out other tasks instead.

BleepingComputer has contacted Google for a comment on this example of Gemini CLI abuse, but we have not received a response as of publishing.

Security teams log 54% of successful attacks and alert on just 14%. The rest move through your environment unseen.

The Picus whitepaper shows how breach and attack simulation tests your SIEM and EDR rules so threats stop slipping by detection.

Advertisement

Get the whitepaper

New buyers will now pay $.50 more per month or $5 a year.

Affordable hybrid IEMs are arriving at a furious pace in 2026, which means another multi-driver in-ear monitor needs more than a padded spec sheet to warrant your attention. SIVGA is making its case at CanJam London 2026 with the Lyrebird, a new $150 quad-driver hybrid IEM built around four drivers and handcrafted stabilized-wood faceplates.

The Lyrebird combines multiple driver types with premium internal components in an effort to deliver a spacious, balanced presentation without wandering into premium-priced territory. At $150, however, it enters a crowded field that includes the $99 ACTIVO Scoop, with its five-driver hybrid array, while DUNU’s four-driver KOTO ITO sits just above it at $199. Both have already demonstrated how much hardware buyers can now expect without emptying their wallets.

The Drivers

The Lyrebird uses a quad-driver hybrid configuration, with each driver assigned to a specific portion of the frequency range. SIVGA says the design is intended to produce a cohesive, natural presentation across the spectrum.

The four-driver array includes:

• A 10mm polymer-composite dynamic driver for low-frequency output, including bass depth, weight, and control.
• A dedicated balanced-armature driver for vocal clarity and detail.
• A micro-planar driver to support the midrange.
• A 9.2mm multilayer piezoelectric ceramic driver for upper-frequency detail and extension.

Craftsmanship

Each Lyrebird earpiece features a handcrafted stabilized-wood faceplate with a unique grain pattern. SIVGA says a specialized resin treatment enhances the appearance of the wood while improving durability and helping to reduce unwanted resonance.

The faceplates are paired with CNC-machined aviation-grade aluminum-alloy housings. The lightweight shells are shaped for an ergonomic fit and are intended to remain comfortable during commuting, home listening, and longer sessions.

Connection Cable

The Lyrebird includes a hybrid cable made with Furukawa oxygen-free copper, silver-plated copper, and gold-plated silver-copper conductors.

A detachable 0.78mm two-pin connection allows the cable to be replaced or upgraded, while the supplied 4.4mm balanced plug is intended for use with compatible headphone amplifiers, digital audio players, and DACs.

Comparison

The Final Touches

The Lyrebird comes with a Crazy Horse leather carrying case and a selection of silicone ear tips to help users achieve a secure fit and proper seal.

With its hybrid driver array, stabilized-wood faceplates, aluminum housings, detachable balanced cable, and included accessories, the Lyrebird arrives as a complete package for home and portable listening.

The Bottom Line 

SIVGA continues to expand its headphone and IEM lineup, and the Lyrebird moves the brand further into affordable territory without abandoning the materials and design details that distinguish its more expensive models.

The Lyrebird is positioned below the 2025 Nightingale Pro, but it is not simply a stripped-down alternative. Its four-driver hybrid array combines a dynamic driver, balanced-armature driver, micro-planar driver, and piezoelectric ceramic driver—an unusually diverse configuration at $150. The stabilized-wood faceplates, aluminum housings, detachable hybrid cable, and 4.4mm balanced termination also give it a stronger visual and technical identity than many competitors at this price.

That price places the Lyrebird in a crowded segment that includes the Fosi Audio IM4 at $99, Sony IER-M500 at $119, and Meze Audio Alba at $159. SIVGA’s combination of handcrafted wood, multiple driver technologies, and balanced connectivity gives the Lyrebird a distinct proposition, although its ultimate value will depend on how successfully those four drivers have been integrated.

The Lyrebird could also create some awkward competition within SIVGA’s own lineup. Its more complex driver configuration may address some of the concerns raised in our Nightingale Pro review, despite costing considerably less, but that remains to be confirmed through listening. Should it perform as intended, the Lyrebird may prove to be more than an affordable entry point into the brand—it could become the SIVGA IEM that offers the strongest balance of design, technology, and price.

Price & Availability

The SIVGA Lyrebird will be first shown and demoed at CanJam London 2026 on July 18-19, 2026.

Related Reading

from the poof-it’s-gone dept

In all of our discussions about how the digital revolution has created a system in which people don’t actually own the things they think they’re buying, I get particularly frustrated by the lack of change in it all. We’ve spilled much ink complaining that this clearly anti-consumer practice needs to be done away with, where an unsuspecting public thinks they’re buying “a thing” only to learn months or years later that “the thing” they bought was actually a license to use/view/listen to another “thing”, and that license exists at the pleasure of the company that collected the money for it. And if you want to see the lack of change or action really honed in upon, let’s take a look at Sony’s PlayStation Store.

In 2022, due to “evolving licensing agreements” with distributor StudioCanal, German and Austrian users had hundreds of movies disappear from their PS accounts, long after buying them through Sony. Then in 2023, it happened again in America, specifically when Sony ended its licensing agreement with Discovery after the Warner Bros. merger, which, of course, has since been bought by Paramount Skydance. That resulted in customers having hundreds and hundreds of episodes of TV shows deleted from their accounts. Nowhere in any of this were there refunds, of course. No recompense at all, actually. Just a thing you thought you’d bought taken away from you by the very people you thought you bought it from.

And now it’s happening again. Due to another licensing agreement fallout with StudioCanal, hundreds of movies and TV shows are being ripped from the accounts of PS Store customers, and there appears to be fuck all that they can do about it.

This news was brought to people’s attention by X user somatyk, who posted the notification they had received from PlayStation this week. Along with the unapologetic news that the purchased movies would be deleted from their account on September 1, the message concluded with, “Click here for a full list of affected titles that will no longer be supported. Thank you.” The same warning is now reproduced in full on the PlayStation website, along with the list of 551 films and TV series that are being pulled from people’s libraries.

As Kotaku notes later in their post, part of what is striking in all of this is the sheer mundanity of the announcement. Because there have been no consequences, or any action at all from the public or government, Sony treats this all as if it’s perfectly normal and no big deal. You can tell me all you want about how the Ts and Cs in these purchases do in fact note that the nature of the purchase is a temporary licensing of the content for an undetermined time period… but I can promise you that the public in general doesn’t understand that. They think they’re buying a thing, not a license.

And that’s because of the purposeful obfuscation of that fact. Sony damned well knows that the vast majority of people don’t read those Ts and Cs. It knows that the public largely doesn’t understand how these backend licensing agreements with distributors work, or that they even exist. And Sony isn’t exactly putting out a big blinking sign on its store pages informing the public of all of this. Instead, the company is only too happy to collect money from a public that is being purposefully kept ignorant of what they’re buying.

Of course, when you scroll past the endless EULAs when you first use your PlayStation, and click “Agree” the first time you load the store, you’re unwittingly agreeing that nothing you buy is really truly bought, and that it can be taken away from you at any point, and there’s nothing you can do. The same is true of your games.

This, too, will probably pass without any real action. The government has done its best to gut our consumer protection agencies, so they won’t be any help. Angry customers won’t coalesce into activism or action, most likely. And I’ll probably be writing another one of these posts in a couple of years when it all happens again.

But it shouldn’t be that way. There are common sense things that can be done to better inform the public. Rules for how the store should inform people with each and every purchase. Someone just needs to demand it be done.

Filed Under: eula, ownership, playstation, playstation store, video games

Companies: sony, studiocanal

It’s likely to become a direct power source for SpaceXAI’s data centers.

Zoom is warning of a critical vulnerability in its desktop client and software development kit for Windows that could be exploited by an unauthenticated party to hijack accounts.

Discovered internally, the security issue is tracked as CVE-2026-53412 and received a severity score of 9.8 out of 10.

In an advisory this week, the messaging platform says that the flaw affects Zoom Workplace for Windows before version 7.0.0, the Windows VDI Client before versions 7.0.10, 6.6.15, and 6.5.18, and the Meeting SDK for Windows before version 7.0.0.

Zoom Workplace, formerly known as Zoom, is a desktop collaboration application for video meetings, group chat, VoIP phone calls, calendar, email, document collaboration, whiteboards, and AI-powered productivity features.

The Windows desktop client is widely deployed and used by millions of individuals and organizations worldwide.

The vendor did not provide any technical details about the flaw in the bulletin, and just described it as an improper input validation issue.

“Improper Input Validation in Zoom Desktop Client for Windows, Zoom VDI Client for Windows, and Zoom Meeting SDK for Windows may allow an unauthenticated user to conduct an account takeover via network access,” reads the security advisory.

To mitigate the risks stemming from CVE-2026-53412, the company recommends that users apply the latest updates.

Zoom’s newest security patches also address the following less severe flaws:

• CVE-2026-53410: high-severity TOCTOU (time-of-check to time-of-use) race condition affecting Zoom Workplace for Windows before 7.0.5, Zoom Workplace VDI Client and VDI Plugin before 6.5.17/6.6.14, Zoom Rooms for Windows before 7.0.5, and Remote Control for Zoom Contact Center before 7.0.0. The flaw could allow an authenticated local user to escalate privileges during installation or uninstallation.
• CVE-2026-53409: high-severity improper privilege management flaw affecting Zoom Rooms for Windows before version 7.1.0 that could allow an authenticated user with local access to escalate privileges.
• CVE-2026-53411: high-severity improper input validation flaw affecting the Zoom Workplace VDI Plugin for Windows before version 6.6.14 that could allow an authenticated user with local access to escalate privileges.

At the time of disclosure, there are no indications that any of the vulnerabilities that Zoom fixed are being exploited in attacks.

Security teams log 54% of successful attacks and alert on just 14%. The rest move through your environment unseen.

The Picus whitepaper shows how breach and attack simulation tests your SIEM and EDR rules so threats stop slipping by detection.

Get the whitepaper

We’ve all had some version of the nightmare: You’ve been inexplicably thrown back into high school, and now you’re standing in front of the class. You’re about to give a speech, but you can’t remember a word. The panic is exacerbated by the faces of students staring at you, some hiding snickers and others openly laughing. Everyone knows what’s going on but you.

I was reminded of this bad dream over the past few years. As my Indigenous school’s instructional technology coordinator, I have seen my office transform from a routine integration hub into a digital confessional. Colleagues and teachers who are trying to navigate destabilizing technology changes slip inside and close the door. Some are looking for practical technology triage, or just a safe space to vent their frustrations, only to recount a similar nightmare. 

For these educators, the experience triggers a sharp wave of panic, a sudden realization that the pedagogical ground has shifted beneath their feet and they can no longer trust their own instincts or traditional technology guardrails. But what follows the panic is a deep, stinging embarrassment. 

With open and hopeful hearts, these teachers publicly praised student writing, extolling the hard work and growth they believed the student had demonstrated, only to find themselves surrounded by the giggles of classmates who already knew the truth: the writing had been fully generated by an AI tool.

The students had technically succeeded while learning very little, a phenomenon Micah Miner describes as unproductive success in “Beyond Secondary Orality.” As a result, teachers and administrators are left trying to respond to tools and behaviors they don’t yet fully understand and have never been trained to address. 

These individual instances of unproductive success were not isolated classroom frustrations; they were early signals of a policy vacuum that would soon overwhelm our entire school system. The challenge becomes infinitely more complex at the school board level, where drafting a formal policy requires balancing the inevitable push toward AI integration with the deeply nuanced and often conflicting perspectives around AI held among students and families in the Native community. Before an AI policy could establish clear lines of accountability for staff and students and safeguard Indigenous data sovereignty, students were already using AI tools outside the school day to engage in cognitive offloading, having essays written for them. 

Because neither an official AI policy nor an academic honesty policy existed, teachers and administrators were forced into the uncomfortable position of responding to what many perceived as cheating without the guidance, definitions or training necessary to do so consistently. 

The problem wasn’t just that students discovered AI before adults did; it was an economy-of-scale issue. Being a one-person technology department meant that the burden of finding a solution and making high-level policy recommendations rested entirely on me. No matter how urgent the crisis, a lone coordinator cannot instantly scale the resources an entire school community needs to decide what AI means, how it should be used, or how we should teach students to navigate it responsibly.

Schools often rely on common sense, hoping that everyone will land on the same understanding. Unfortunately for us, AI has exposed an uncomfortable truth: common sense is often just unspoken assumptions masquerading as policy. Without a common understanding of AI use, implementation practices, potential risks and benefits, schools risk creating inconsistent and inequitable experiences for both students and educators. This raises the question: who decides what is reasonable or what makes a decision good? 

Answering that question didn’t require finding a perfect, pre-packaged solution. Ultimately, it required a small but dedicated team of school leaders, educators and me to build one from the ground up. 

Building a Framework on Common Sense and Community 

For the past several years, it has been my job to learn as much as I can to create our own AI common sense, one based on the mission of our school to prepare our students to carry forward the wisdom of our ancestors for the benefit of future generations. 

My first step was to seek out as much information as possible from trusted sources. This led me to the International Society for Transforming Education and their AI Explorations course, which established my foundational knowledge around AI and machine learning and created a strong network of other professionals who were also trying to navigate without a map. 

My second step was to reach out to my school’s stakeholders to ensure the work on AI benefited from as many diverse perspectives as possible. As a new instructional technology coordinator in 2017, one of the first things I did when establishing my program at Indian Community School was to create a Future Ready Team, a cross-functional group of teachers, IT and administrators. 

These Future Ready Team members became the resident experts during their two-year terms, learning to communicate with vendors and piloting new software and devices. When the issue of AI policy arose, it was this team that rose to the challenge and continued the research process along with me.  

After months of research and collaboration with the Future Ready Team, my binder was overflowing with highlighted articles, ideas, notes and guidance documents — proof that some habits from my 1990s school days die hard — while my computer held an equally overwhelming collection of bookmarked tabs. 

Faced with this data overload, I knew my job was to translate this research into action. The solid footing we needed was a comprehensive AI framework, so I took the lead on designing a model that would serve as a strong foundation for building knowledge and be flexible enough to adapt to the ever-changing AI landscape. 

I personally curated and authored each layer of this framework to meet our school’s specific needs. This included a core policy I developed using resources from the U.S. Department of Education, UNESCO, the Publications office of the European Union and the Wisconsin Department of Public Instruction. I also drafted staff AI guidance based on the AI Guidance for Schools Toolkit by TeachAI, revamped our student responsible use policy to incorporate AI guardrails and constructed a visual stoplight of AI usage for classrooms based on the work done in our home state of North Carolina. 

The framework also needed to do more than establish expectations; it needed to support educators in meeting them by addressing critical issues such as safeguarding personally identifiable information, protecting intellectual property and honoring Indigenous data sovereignty. 

Policies Can’t Teach People

While these protocols were a necessary first step in securing our digital environment, a framework is only as effective as the people implementing it. Creating the policy and guidance documents was only part of the work; true equity emerges in a school community’s capacity to respond, specifically having the staff, training, time and funding needed to translate policy into practice. 

Knowing that even the best guidance will sit unused without these resources, we ensured our framework went beyond rules to provide active opportunities for people to learn, question and build understanding together. 

Before students were given access to AI tools on school devices, staff engaged in professional development designed to build AI literacy and confidence. Learning opportunities continued through local and state conferences, while families and community members were invited to join the conversation at an AI literacy night. The goal was not simply to introduce a new technology, but to build a shared foundation for its responsible use. The outcomes of this collective effort extended far beyond compliance. 

Teachers who had previously felt overwhelmed by the sudden influx of these tools gained the confidence needed to navigate them, shifting their approach from policing technology to guiding its purposeful use. 

Similarly, our engagement with families provided parents and community members with a transparent look at what AI usage actually looks like in classrooms. This shared clarity transformed potential anxiety into partnership, ensuring the entire school community was aligned around a unified approach that protects student critical thinking while fostering academic readiness. 

Enduring Understandings Benefit All

Ultimately, this journey has taught me that AI common sense isn’t something we can simply mandate or print out. It isn’t found neatly tucked inside a binder, no matter how many national frameworks we compile. True common sense has to be built intentionally through the collaborative and sometimes uncomfortable work of research, reflection and learning alongside one another. 

While a policy can draw the lines on a map, it is the shared knowledge and trust within a school community that actually gives us the solid footing to move forward. 

At Indian Community School, that forward path is always guided by the principle of thinking seven generations ahead. It is a grounding reminder that the choices we make today about AI are not just about managing next week’s essays or updating a student handbook. They are about shaping students’ learning habits and critical thinking abilities long after they leave us. 

For schools everywhere, the challenge now is to look past the establishment of compliance. We must close our binders, sit down with our communities and commit to the slower, more deliberate work of building shared understanding around AI. We have to stop racing to build rules and instead start investing in the training, time and relationships that ground our educators for whatever comes next.