Shuttered Startups Are Selling Old Slack Chats, Emails To AI Companies
- Reference: 0181742456
- News link: https://yro.slashdot.org/story/26/04/18/014244/shuttered-startups-are-selling-old-slack-chats-emails-to-ai-companies
- Source link:
> Shanna Johnson, the CEO of now-defunct software company Cielo24, told the publication that she was able to sell every Slack message, internal email, and Jira ticket as training data for "hundreds of thousands of dollars."
>
> This isn't a one-off scenario. SimpleClosure, a startup that helps companies like Cielo24 shut down, told Forbes that there's been major interest from AI companies trying to get their hands on workplace data. Because of this, SimpleClosure launched a new tool that allows companies to sell their wealth of internal communications -- from Slack archives to email chains -- to AI labs. The company said it's processed 100 such deals in the past year. Payouts ranged from $10,000 to $100,000.
"I think the privacy issues here are quite substantial," Marc Rotenberg, 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. It's identifiable people."
[1] https://www.forbes.com/sites/annatong/2026/04/16/ais-new-training-data-your-old-work-slacks-and-emails/
[2] https://www.fastcompany.com/91528808/shuttered-startups-are-selling-old-slack-chats-and-emails-to-ai-companies
No privacy was expected in the workplace but ... (Score:2)
Never expected privacy in the workplace, but never anticipated this.
Wonder a treasure trove of slackness will leak out from one of these models.
Interesting strategy.. (Score:5, Insightful)
Train models mostly on the communications of failed businesses... What could go wrong?
Re: (Score:2)
Great concept. If I owned a failing business, I'd use an LLM to create fake communication records so I could sell them to AI companies. Shouldn't be too difficult to create a fake message trail, then run a script that checks basic time ordering of timestamps and dates.
Interesting future. (Score:2)
> Great concept. If I owned a failing business, I'd use an LLM to create fake communication records so I could sell them to AI companies. Shouldn't be too difficult to create a fake message trail, then run a script that checks basic time ordering of timestamps and dates.
It's gonna be an ironic bitch when AI finds the real company went bankrupt and closed, but the fake company they were running on an LLM filed for an IPO in the Metasphere after making $100 million last year writing fake communication records for bots pretending to be meatsack workers.
Re: (Score:2)
That is an excellent point. There is a lot going on with training data that the AI sellers do not seem to understand. In their haste (and clearly panic) to make their products seem to be absolutely great and something they are not, they seem to stop at almost nothing now.
Re: (Score:2)
> Train models mostly on the communications of failed businesses... What could go wrong?
80% of new businesses fail within the first 3 years.
If you were starting a new business, would you NOT want to know all of the ways to help mitigate an 80% chance of failure?
Imagine I asked you that about marriage..
at least someone wants it (Score:3)
I've seen more than one San Diego startup bite the dust. I recall once (was late 90's, biotech layoffs were the norm) going back to the building because I'd heard that the company was auctioning office stuff. When I rolled up, I saw that a 2nd story window had been popped out. There was a group of workers chucking boxes of papers and lab notebooks into a dumpster down below.
GDPR (Score:4, Insightful)
This cannot be legal in the EU, right? Or even if only one of your employees works remotely from EU?
Re:GDPR (Score:4, Informative)
It is not, unless all private or chat use was prohibited. That is difficult. because it is perfectly acceptable to use the company email to mail home "sorry, I am going to work an hour longer" or "lets have lunch at xyz today" to a co-worker. And those messages belongs to the user and are protected under the GDPR and cannot be used in any new ways the user did not consent to.
Re: GDPR (Score:2)
Nah, this thread is onto something. I've seen policies to scrub old employee info from all systems within X amount of years after they leave.
The circle, darkens. (Score:2)
> And when the bubble pops AI companies will have to sell their assholes too.
I fix, er edited that for you. On behalf of the immoral filter apparently ruining my quick-scan vision in the millionaire OF world we live in.
Ex-AI employees booty-grinding on a bartop in Borneo, bending over to make that butthole bitcoin? Damn. And I thought the .bomb bubble pop was sad.
A serious question (Score:2)
There seems to be some potential for a lot of good to come out of AI, but right now it seems to be putting its weight behind civilization's slide into a dystopian hellscape.
Never mind will we - rather, can we - turn the tide and make AI work for the good of all humanity, or at least make it neutral?
Is AI so dangerous that it can never support the overall common good? Or are we so fucked up that we are unable to make it work for the collective good?
Sorry, that was more than one serious question - my bad.
Wouldn't it be nice to have a GDPR? (Score:4, Interesting)
Because then this would be completely illegal.
Dang (Score:2)
Sure, it might be legal. But only because in a functioning society you normally don't need to legislate everything.
New business opportunity (Score:2)
1. Create a fake startup
2. Generate mountains of plausible-looking emails and chat messages using an LLM.
3. Close the startup and... Profit.
Re: (Score:2)
Sounds like an easy side hustle if you set aside enough hours to get it going. This could be your winning lottery ticket idea and no one of importance would be defrauded. I promise I'll side with you if you ever have to answer to the court system. Nope, he's innocent.
Slack's been doing this for years (Score:2)
As in: [1]Slack users horrified to discover messages used for AI training [arstechnica.com]
This should surprise nobody. Slack is one of the sleaziest companies on the planet, and there's a lot of competition for distinction.
[1] https://arstechnica.com/tech-policy/2024/05/slack-defends-default-opt-in-for-ai-training-on-chats-amid-user-outrage/
Employee conversation in work environment (Score:2)
Are not exactly private, as your HR will make it very clear. But I do hope someone spends the time to adequately review all those chats and emails to make sure no private client information leaks out... Right?
Re: Employee conversation in work environment (Score:2)
Imagine the possibilities. 1) Hmm, well Bob you bitched about the boss and coworkers a lot in your last three startups. Weâ(TM)re going to pass on hiring you. 2) looks like you code well but your former coworkers complained you were lazy. 3) we built a profile on you from former and current messaging and it's clear you are coasting at this point until retirement.. your fired. I guess the last one might be good for long term employees at any company
Give-a-shit, is rare after the fact. (Score:2)
> Are not exactly private, as your HR will make it very clear. But I do hope someone spends the time to adequately review all those chats and emails to make sure no private client information leaks out... Right?
Quite a few years ago, I was called up for jury duty on a Federal case involving a restaurant chain and several accusations of sexual misconduct and harassment against a manager. Probably half a dozen women involved in the case, with this taking place years after the fact. Meaning the people involved had established their lives in society with jobs, relationships, loans, credit, etc.
The defense brought in half a dozen volumes of evidence in 3" binders. Time cards, employee records, all manner of document