Goto

Collaborating Authors

 contractor



OpenAI Is Asking Contractors to Upload Work From Past Jobs to Evaluate the Performance of AI Agents

WIRED

To prepare AI agents for office work, the company is asking contractors to upload projects from past jobs, leaving it to them to strip out confidential and personally identifiable information. OpenAI is asking third-party contractors to upload real assignments and tasks from their current or previous workplaces so that it can use the data to evaluate the performance of its next-generation AI models, according to records from OpenAI and the training data company Handshake AI obtained by WIRED. The project appears to be part of OpenAI's efforts to establish a human baseline for different tasks that can then be compared with AI models. In September, the company launched a new evaluation process to measure the performance of its AI models against human professionals across a variety of industries. OpenAI says this is a key indicator of its progress towards achieving AGI, or an AI system that outperforms humans at most economically valuable tasks. "We've hired folks across occupations to help collect real-world tasks modeled off those you've done in your full-time jobs, so we can measure how well AI models perform on those tasks," reads one confidential document from OpenAI.


The US Military Wants to Fix Its Own Equipment. Defense Contractors Are Trying to Shoot That Down

WIRED

A push by military contractors could alter pending legislation that would have empowered servicemembers to repair equipment. Lobbyists are pitching a subscription service instead. Right to repair provisions in the National Defense Authorization Act, which would secure funding for the US military in 2026, are likely to be struck from the final language of the bill despite enjoying broad bipartisan support, sources familiar with ongoing negotiations tell WIRED. They say that provisions in the act enabling servicemembers to repair their own equipment are likely to be removed entirely, and replaced with a data-as-a-service subscription plan that benefits defense contractors. The right to repair has become a thorny issue in the military.


Supplementary Information A Collecting Internet Data A.1 Initial Unclean Dataset Curation

Neural Information Processing Systems

Our goal was to curate a video dataset of Minecraft gameplay from the survival game mode. Minecraft Survival game mode that include such visual artifacts. Please help us identify screenshots that belong only to the survival mode in Minecraft. Survival mode is identified by the info at the bottom of the screen: a health bar (row of hearts) a hunger bar (row of chicken drumsticks) a bar showing items held Survival Mode V alid survival mode videos have health/hunger bars and an item hotbar at the bottom of the screen. Creative Mode Creative mode only has an item hotbar and should be classified as None of the Above .


Adversarial training for high-stakes reliability Daniel M. Ziegler Seraphina Nix Lawrence Chan

Neural Information Processing Systems

In the future, powerful AI systems may be deployed in high-stakes settings, where a single failure could be catastrophic. One technique for improving AI safety in high-stakes settings is adversarial training, which uses an adversary to generate examples to train on in order to achieve better worst-case performance. In this work, we used a safe language generation task ("avoid injuries") as a testbed for achieving high reliability through adversarial training. We created a series of adversarial training techniques--including a tool that assists human adversaries--to find and eliminate failures in a classifier that filters text completions suggested by a generator. In our task, we determined that we can set very conservative classifier thresholds without significantly impacting the quality of the filtered outputs. We found that adversarial training increased robustness to the adversarial attacks that we trained on--doubling the time for our contractors to find adversarial examples both with our tool (from 13 to 26 minutes) and without (from 20 to 44 minutes)--without affecting in-distribution performance. We hope to see further work in the high-stakes reliability setting, including more powerful tools for enhancing human adversaries and better ways to measure high levels of reliability, until we can confidently rule out the possibility of catastrophic deployment-time failures of powerful models.


Meta Claims Downloaded Porn at Center of AI Lawsuit Was for 'Personal Use'

WIRED

Meta Claims Downloaded Porn at Center of AI Lawsuit Was for'Personal Use' In a motion to dismiss filed earlier this week, Meta denied claims that employees had downloaded pornography from Strike 3 Holdings to train its artificial intelligence models. This week, Meta asked a US district court to toss a lawsuit alleging that the tech giant illegally torrented pornography to train AI . The move comes after Strike 3 Holdings discovered illegal downloads of some of its adult films on Meta corporate IP addresses, as well as other downloads that Meta allegedly concealed using a "stealth network" of 2,500 "hidden IP addresses." Accusing Meta of stealing porn to secretly train an unannounced adult version of its AI model powering Movie Gen, Strike 3 sought damages that could have exceeded $350 million, TorrentFreak reported . Strike 3 also cited "no facts to suggest that Meta has ever trained an AI model on adult images or video, much less intentionally so," Meta claimed.


China Dives in on the World's First Wind-Powered Undersea Data Center

WIRED

The $226 million project uses ocean breezes and seawater to stay cool. China is submerging data centers into the ocean to keep them cool. China has completed the first phase of construction of what it claims is the world's first underwater data center (UDC). Located in Shanghai's Lin-gang Special Area with a price tag of roughly RMB 1.6 billion ($226 million), it's a significant milestone in the quest for sustainable solutions to the growing energy demands of China's computing infrastructure. Powered entirely by wind energy, the initiative has a total power capacity of 24 megawatts.


A Plan to Rebuild Gaza Lists Nearly 30 Companies. Many Say They're Not Involved

WIRED

Many Say They're Not Involved A presentation that has been shared with the Trump administration references Tesla, Ikea, TSMC, and more in its plan to rebuild Gaza. Some of these companies say they had no idea they were mentioned. The mound of rubble at the site of the Unknown Soldier Tower, destroyed by overnight Israeli bombardment, is pictured in the Rimal neighbourhood of Gaza City on September 15, 2025. A sweeping plan to reconstruct Gaza, which has been shared with Trump administration officials, features the names and logos of more than two dozen companies--some of which tell WIRED they had no knowledge they were named or involved. The presentation outlining the plan was reportedly created by some of the businessmen who helped ideate what became the controversial nonprofit the Gaza Humanitarian Foundation, which is currently leading aid distribution in Gaza, calling for the creation of a new entity called the Gaza Reconstitution, Economic Acceleration and Transformation (GREAT) Trust.



Palantir Wants to Be a Lifestyle Brand

WIRED

Defense tech giant Palantir is selling T-shirts and tote bags as part of a bid to encourage fans to publicly endorse it. Palantir Technologies, which moved from Silicon Valley to Denver in 2020, sells software that immigration authorities use to identify and arrest people, militaries use to organize drone strikes, and corporations use to manage their supply chains. Now, it also sells tote bags. Last year, Palantir re launched an online merchandise store, and its website was recently redesigned with a swanky interface and new payment system . A mock terminal in the lower left corner displays "code" documenting each item you view.