This is pretty scary:

Urban VPN Proxy targets conversations across ten AI platforms: ChatGPT, Claude, Gemini, Microsoft Copilot, Perplexity, DeepSeek, Grok (xAI), Meta AI.

For each platform, the extension includes a dedicated “executor” script designed to intercept and capture conversations. The harvesting is enabled by default through hardcoded flags in the extension’s configuration.

There is no user-facing toggle to disable this. The only way to stop the data collection is to uninstall the extension entirely.

[…]

The data collection operates independently of the VPN functionality. Whether the VPN is connected or not, the harvesting runs continuously in the background...

States with booming data center construction are seeing spikes in new power needs. Much of it is being met by coal and solar.

Technology firms and their trade groups may not be doing enough to sway an increasingly skeptical public as electricity prices climb.

The year featured unprecedented assaults on state climate action. Democrats also think it showed them a path back to power.

A yearlong assault on federal science programs taught researchers they can't always count on Washington to be a reliable partner.

Nearly $2.8 billion is set for state projects ranging from hybrid ferries to bicycle rebates.

Proposed data centers in the state would require huge amounts of power while likely bringing few long-term jobs.

Campers will start arriving in May, bunking on higher ground than the area where fast-rising waters swept away two cabins this year.

A reliance on Europe’s ERA5 runs counter to Beijing’s push for security and technological independence.

When customers choose earlier delivery dates, the system shifts from optimized routing to getting the package out fastest.

Nature Climate Change, Published online: 24 December 2025; doi:10.1038/s41558-025-02532-6

Heatwave attribution in seconds

Nature Climate Change, Published online: 24 December 2025; doi:10.1038/s41558-025-02530-8

Foraging constrained by heat and dark

Nature Climate Change, Published online: 24 December 2025; doi:10.1038/s41558-025-02529-1

Rising lake and reservoir emissions

Nature Climate Change, Published online: 24 December 2025; doi:10.1038/s41558-025-02531-7

Inequalities in resilience and preparedness

Nature Climate Change, Published online: 24 December 2025; doi:10.1038/s41558-025-02508-6

Whether erosion is accelerating or decelerating along Arctic rivers has been unclear, but each trend has distinct implications for the vast amount of carbon stored in permanently frozen soils. Now, research demonstrates that warming air temperatures are driving divergent outcomes for Arctic rivers, causing some to erode their banks more rapidly while others slow down.

Nature Climate Change, Published online: 24 December 2025; doi:10.1038/s41558-025-02525-5

Climate change drives displacement and migration across the Americas, particularly exposing Latin American and Caribbean children to compounded health risks. We explore these health impacts, identify gaps in related US healthcare and health policy, and propose recommendations for how they can respond.

Nature Climate Change, Published online: 24 December 2025; doi:10.1038/s41558-025-02512-w

Whether rivers are speeding up or slowing down in a warming Arctic is unclear, but has implications for carbon cycling and infrastructure. This study finds divergent behaviour in migration rates for rivers in discontinuous versus continuous permafrost, driven by changes in permafrost thaw and river ice.

Nature Climate Change, Published online: 24 December 2025; doi:10.1038/s41558-025-02520-w

Energy transition minerals (ETM) are essential for decarbonization, yet extractions often occur in carbon-rich forests and lands of Indigenous peoples and local communities. Here the authors provide global analysis showing how ETM mining causes sustained forest loss and GHG emissions.

The world has been forced to bear the weight of billionaires and politicians who salivate over making tech more invasive, more controlling, and more hostile. That's why EFF’s mission for your digital rights is crucial, and why your support matters more than ever. You can fuel the fight for privacy and free speech with as little as $5 or $10 a month:

Join EFF

Become a Monthly Sustaining Donor

When you donate by December 31, your monthly support goes even further by unlocking bonus Year-End Challenge grants! With your help, EFF can receive up to seven grants that increase in size as the number of supporters grows (check our progress on the counter). Many thanks to EFF’s Board of Directors for creating the 2025 challenge fund.

The EFF team makes every dollar count. EFF members giving just $10 or less each month raised $400,000 for digital rights in the last year. That funds court motions, software development, educational campaigns, and investigations for the public good every day. EFF member support matters, and we need you.

📣 Stand Together: That’s How We Win 📣

You can help EFF hold corporations and authoritarians to account. We fight for tech users in the courts and we lobby and educate lawmakers, all while developing free privacy-enhancing tech and educational resources so people can protect themselves now. Your monthly donation will keep us going strong in this pivotal moment.

Get your choice of free gear when you join EFF!

Your privacy online and the right to express yourself are powerful—and it’s the reason authoritarians work so viciously to take them away. But together, we can make sure technology remains a tool for the people. Become a monthly Sustaining Donor or give a one-time donation of any size by December 31 and unlock additional Year-End Challenge grants!

Give Today

Unlock Year-End Challenge Grants

Already an EFF Member? Help Us Spread the Word!

EFF Members have carried the movement for privacy and free expression for decades. You can help move the mission even further! Here’s some sample language that you can share with your networks:

We need to stand together and ensure technology works for us, not against us. Donate any amount to EFF by Dec 31, and you'll help unlock challenge grants! https://eff.org/yec
Bluesky | Facebook | LinkedIn | Mastodon
(more at eff.org/social)

_________________

EFF is a member-supported U.S. 501(c)(3) organization. We’re celebrating TWELVE YEARS of top ratings from the nonprofit watchdog Charity Navigator! Your donation is tax-deductible as allowed by law.

In 2024, EFF wrote our initial blog about what could go wrong when police let AI write police reports. Since then, the technology has proliferated at a disturbing rate. Why? The most popular generative AI tool for writing police reports is Axon’s Draft One, and Axon also happens to be the largest provider of body-worn cameras to police departments in the United States. As we’ve written, companies are increasingly bundling their products to make it easier for police to buy more technology than they may need or that the public feels comfortable with. 

We have good news and bad news. 

Here’s the bad news: AI written police reports are still unproven, untransparent, and downright irresponsible–especially when the criminal justice system, informed by police reports, is deciding people’s freedom. The King County prosecuting attorney’s office in Washington state barred police from using AI to write police reports. As their memo read, “We do not fear advances in technology – but we do have legitimate concerns about some of the products on the market now... AI continues to develop and we are hopeful that we will reach a point in the near future where these reports can be relied on. For now, our office has made the decision not to accept any police narratives that were produced with the assistance of AI.” 

In July of this year, EFF published a two-part report on how Axon designed Draft One to defy transparency. Police upload their body-worn camera’s audio into the system, the system generates a report that the officer is expected to edit, and then the officer exports the report. But when they do that, Draft One erases the initial draft, and with it any evidence of what portions of the report were written by AI and what portions were written by an officer. That means that if an officer is caught lying on the stand – as shown by a contradiction between their courtroom testimony and their earlier police report – they could point to the contradictory parts of their report and say, “the AI wrote that.” Draft One is designed to make it hard to disprove that. 

In this video of a roundtable discussion about Draft One, Axon’s senior principal product manager for generative AI is asked (at the 49:47 mark) whether or not it’s possible to see after-the-fact which parts of the report were suggested by the AI and which were edited by the officer. His response (bold and definition of RMS added): 

“So we don’t store the original draft and that’s by design and that’s really because the last thing we want to do is create more disclosure headaches for our customers and our attorney’s offices—so basically the officer generates that draft, they make their edits, if they submit it into our Axon records system then that’s the only place we store it, if they copy and paste it into their third-party RMS [records management system] system as soon as they’re done with that and close their browser tab, it’s gone. It’s actually never stored in the cloud at all so you don’t have to worry about extra copies floating around.”

Yikes! 

All of this obfuscation also makes it incredibly hard for people outside police departments to figure out if their city’s officers are using AI to write reports–and even harder to use public records requests to audit just those reports. That’s why this year EFF also put out a comprehensive guide to help the public make their records requests as tailored as possible to learn about AI-generated reports. 

Ok, now here’s the good news: People who believe AI-written police reports are irresponsible and potentially harmful to the public are fighting back. 

This year, two states have passed bills that are an important first step in reigning in AI police reports. Utah’s SB 180 mandates that police reports created in whole or in part by generative AI have a disclaimer that the report contains content generated by AI. It also requires officers to certify that they checked the report for accuracy. California’s SB 524 went even further. It requires police to disclose, on the report, if it was used to fully or in part author a police report. Further, it bans vendors from selling or sharing the information a police agency provided to the AI. The bill also requires departments to retain the first draft of the report so that judges, defense attorneys, or auditors could readily see which portions of the final report were written by the officer and which portions were written by the computer.

In the coming year, anticipate many more states joining California and Utah in regulating, or perhaps even banning, police from using AI to write their reports. 

This article is part of our Year in Review series. Read other articles about the fight for digital rights in 2025.