惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

美团技术团队
D
DataBreaches.Net
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
D
Docker
N
Netflix TechBlog - Medium
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
C
Check Point Blog
腾讯CDC
Stack Overflow Blog
Stack Overflow Blog
V
Visual Studio Blog
IT之家
IT之家
月光博客
月光博客
U
Unit 42
K
Kaspersky official blog
T
Threatpost
cs.AI updates on arXiv.org
cs.AI updates on arXiv.org
GbyAI
GbyAI
P
Proofpoint News Feed
Last Week in AI
Last Week in AI
云风的 BLOG
云风的 BLOG
酷 壳 – CoolShell
酷 壳 – CoolShell
I
InfoQ
Engineering at Meta
Engineering at Meta
Recorded Future
Recorded Future
Exploit-DB.com RSS Feed
Exploit-DB.com RSS Feed
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
S
Security @ Cisco Blogs
MyScale Blog
MyScale Blog
大猫的无限游戏
大猫的无限游戏
Security Archives - TechRepublic
Security Archives - TechRepublic
Webroot Blog
Webroot Blog
cs.CV updates on arXiv.org
cs.CV updates on arXiv.org
Hacker News - Newest:
Hacker News - Newest: "LLM"
S
Schneier on Security
S
Secure Thoughts
The Register - Security
The Register - Security
B
Blog RSS Feed
The Last Watchdog
The Last Watchdog
P
Palo Alto Networks Blog
爱范儿
爱范儿
B
Blog
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
N
News and Events Feed by Topic
阮一峰的网络日志
阮一峰的网络日志
L
LINUX DO - 热门话题
C
Cisco Blogs
Spread Privacy
Spread Privacy
F
Full Disclosure
博客园 - 聂微东
T
The Blog of Author Tim Ferriss

Futurism

Meta’s AI Support Bot Is Giving Hackers Access to Other People’s Instagram Accounts Just by Asking Websites Are Spying on Your Solid State Drive The Trump Phone Appears to Have Already Leaked Its Customers’ Personal Information Through a Glaring Exploit Vibe Coded Apps Are Spilling Users’ Personal Information Directly Into the Maw of Greedy Hackers How to Get Rid of Reddit’s Giant App-Shilling Popup That Breaks Its Entire Mobile Site Madison Square Garden Reportedly Used Facial Recognition to Stalk Trans Woman For Two Years Top Security Experts Alarmed by Power of Anthropic’s New Hacker AI Companies Just Learned a Brutal Lesson About Training AI to Do Human Jobs Huge Group of Experts Warns Meta That Its Pervert Glasses Will Enable Terrible Crimes
Your Former Employer Is Selling Your Slacks and Emails to Train AI
Frank Landym · 2026-04-25 · via Futurism

Two hands exchanging a stack of US 100-dollar bills, with one hand giving and the other receiving the money. The background is dark, and the hands are in focus, highlighting the transaction.

Getty / Futurism

Sign up to see the future, today

Can’t-miss innovations from the bleeding edge of science and tech

There’s a new link in the food chain of tech startups. Founders are realizing that they can make an extra buck by selling off the digital remains of their dead companies in the form of their employees’ Slack messages and emails. And now, there’s a whole ecosystem of decomposers that specialize in helping them make that happen, Forbes reports.

Getting quality training data is half the battle with building an AI model, and fresh sources are increasingly hard to come by in a post-AI-ransacked internet. The data derived from a digital workplace like Slack is particularly valuable because of the industry’s heavy focus on building AI agents that can carry out work tasks, allowing AI researchers to construct a realistic sandbox the agents can learn in.

“Model companies are realizing the noise in the real-world environments is required to accurately test models,” Ali Ansari, whose company micro1 sells a mock holding company for AI agents to learn the ropes in.

These simulated workplace sandboxes are being called reinforcement learning gyms, or “RL gyms,” and creating them has become its own nascent industry. The demand is such that Anthropic is considering spending $1 billion on RL gyms this year, The Information reported, and multiple RL gym startups, such as Prime Intellect and Fleet, are now being valued around that same figure.

Founders of defunct startups are more than happy to do business with RL gym builders, and middlemen have emerged to facilitate these transactions. SimpleClosure, a company that styles itself the “TurboTax of shutting down,” recently launched a new tool called Asset Hub that allows moribund companies to sell off their Slack archives, emails, and libraries of code, and supposedly anonymizing that data before it finds a buyer. Its CEO Dori Yona told Forbes that it’s processed nearly 100 deals for defunct companies in the past year, recovering over $1 million dollars for their founders.

Of course, there’s an undeniable tension at the heart of the phenomenon. If you want to train an AI to succeed in business, is training it on the practices of a company that went out of business the best preparation?

And wherever data’s concerned, there are clear ethical and privacy risks. It would be an outrage if an employer recorded every conversation their employees had in a physical workplace and then sold them for profit, but when those conversations are facilitated digitally, selling them is seen as a shrewd bit of business that’s helping build the next generation of AI.

“I think the privacy issues here are quite substantial,” Marc Roteberg, founder of the Center for AI and Digital Policy, told Forbes. “Employee privacy remains a key concern, particularly because people have become so dependent on these new internal messaging tools like Slack.” 

“It’s not generic data,” Roteberg added. “It’s identifiable people.”

Experts also contest claims that the companies are anonymizing the data they’re pawning off.

“If anonymization’s not done correctly, there are risks that companies who have access to the data would be able to see the activities of individual organizations and people, and then if not treated carefully, could leak into model output,” Bobby Samuels, the CEO of Protege, a company that vets and sells data to AI developers, told Forbes.

More on AI: JPMorganChase Data Center Gets $77 Million Handout to Create Grand Total of One Job