- 11 hours ago
Category
🗞
NewsTranscript
00:00I hate everything Meta has done.
00:02I hate that I work for psychopaths.
00:04I despise that I'm part of the machine.
00:07Workers inside big tech companies
00:09are getting increasingly disillusioned
00:11with the world they helped build.
00:13I can't keep selling my soul like this.
00:15This is a prison in its own way.
00:18Tech workers and CEOs are increasingly worried
00:21that AI is going to kill us all.
00:22But this doomerism is obscuring
00:25what AI is actually doing to workers right now.
00:28So I spoke to over a dozen tech workers
00:30to hear what it's like on the inside.
00:33These are not their real voices.
00:35Zuck thinking that we could replace 60% of the company
00:38with AI agents?
00:40It's like you're insane.
00:41It doesn't matter how much money you make for them.
00:43They will turn on you.
00:45And the more people I spoke to,
00:47the more I realized something weirder was going on.
00:50We didn't really know how,
00:51but I knew somebody was watching.
00:54This thing records your screen, records clicks.
00:56It just seems like their goal is to automate us away.
01:01Tech workers aren't the only ones being watched.
01:03About two-thirds of all workers in the U.S.
01:06say they are being surveilled by technology known as Vosware.
01:09And now it's getting supercharged by AI.
01:13Here at Amazon, they only treat us by the metrics and numbers.
01:16They don't give a about you as a human.
01:18So I wanted to know, how is this AI-powered surveillance technology
01:22affecting workers today?
01:24Is all this worker data going to train our AI replacements?
01:28And what does Vosware think of me?
01:32Vosware has been around since people like Henry Ford
01:35surveilled and controlled workers every move.
01:37Even though the technology has changed a lot over the years,
01:40the directive is still the same – collect as much data as possible
01:45to maximize worker productivity and profits at all costs.
01:50Everything we do, everything we say, everything is surveilled.
01:55Sherry Shaw has been a nurse for 28 years
01:58and currently answers triage calls for Kaiser Permanente.
02:02We're watched for our call times.
02:04We'll watch how fast we answer the phone.
02:06If we've got to get off the phone to do anything personal,
02:09that's also monitored.
02:11Sherry and her fellow nurses are only supposed to be on calls
02:14with patients for 13 minutes.
02:16We've asked multiple times, where did you come up with 13 minutes?
02:20And then they say, oh, well, studies show.
02:22I'm like, well, can you show us the studies?
02:24I mean, what are you talking about?
02:26Sherry says that she and other nurses can be disciplined
02:29when they don't hit certain targets.
02:31However, Kaiser told us that they don't use these metrics
02:34to assess performance.
02:36If they're worried about losing their job,
02:38they're going to try to work faster.
02:39And when they work faster, that's when they miss something.
02:42That's when patient care becomes unsafe.
02:44Studies have shown that boss work can make working conditions
02:47less safe and less humane, like Amazon drivers who pee in bottles
02:52and poop in doggy bags just to stay on schedule.
02:55And now with the rise of AI, this technology is getting more invasive
02:59for blue-collar and service jobs.
03:02In warehouses, algorithms can fire employees.
03:05In Amazon delivery trucks, AI-powered cameras flag when a driver yawns.
03:10In Burger King, an AI assistant called Patty is in employees' ears,
03:14telling them what to do.
03:15Looks like the women's restroom needs cleaning.
03:18That's very helpful, Patty. I'll get to it right away.
03:21For call center workers, Emotion AI can judge how empathetic workers are.
03:26That's what happened to Sherry Shaw,
03:28who was monitored by an Emotion AI pilot program.
03:31I was like forever being told by this AI that my tone was wrong,
03:37that my voice didn't sound soft, it didn't sound engaging,
03:41it didn't sound like I cared.
03:43We're not robots, but they want us to act like robots.
03:48To truly understand how these systems work,
03:50I made a somewhat dubious decision.
03:53I started surveilling myself.
03:55I downloaded Bossware software that got unleashed on remote workers
03:58during the pandemic and has persisted today.
04:01I could click in to see exactly what's on my computer in real time.
04:05There were photos of me staring blankly at my screen.
04:08It kept alerting me that I violated arbitrary rules.
04:11I got insights into how I could be more productive
04:14or what my company-wide sentiment was.
04:16When I started digging into the metrics,
04:19I noticed some of my supposed unproductive time was for things that are legitimately part of my job,
04:24like reading articles about Bossware, watching call center webinars, and making calls.
04:31Some Bossware companies use all this data to generate a productivity score,
04:35which can have real consequences.
04:37So I signed up for a product demo to get an insider look at the back end of these scores.
04:42How do you determine who's a low performer?
04:46What I would recommend looking at is the productivity trends.
04:49So that's based on the productivity score then?
04:52Yes.
04:53Since we're using company laptops,
04:55would they know that this software is on their computer?
05:00That's optional.
05:02Oftentimes, employees don't know the extent that they're getting surveilled.
05:05But since I was surveilling myself, I checked in on one of my productivity scores,
05:10and it was a dismal 65 percent.
05:13Sorry, more perfect union.
05:14Many people don't know what's going into a worker productivity score.
05:18So it's moving at a speed and a scale and a level of opacity that is concerning.
05:25Samantha Gordon has been following trends in Bossware for years.
05:28This moment in Bossware, whether you're at headquarters or you're in a warehouse
05:33or you're a gig driver, folks understand more and more and more
05:37that our fates are linked together.
05:39For years, these opaque scores have had real consequences on low-wage and gig workers
05:44like Uber drivers.
05:46Now, these scores are increasingly coming for more and more people,
05:50including tech workers themselves.
05:52A recent lawsuit alleges that Meta used AI-assisted tools to score, rank,
05:57and suggest who should be laid off.
05:59I've seen cases where people have gotten, like, great reviews year after year.
06:04And then one year, they take medical leave and all of a sudden,
06:07they get a bad rating and they're gone.
06:09The lawsuit alleges that AI-assisted tools
06:12discriminated against people who took parental or medical leaves
06:16because it scored them as less productive.
06:18In a statement to More Perfect Union, Meta said that organizational decisions
06:22were and are made by people, not AI.
06:25I sort of suspected that I was going to be on the list.
06:29Since joining Meta, I've gained, like, 50 pounds in a year and a half,
06:33had panic attacks, like, really, really bad ones,
06:36where it's like we had to check out whether I had, like, some brain damage.
06:40And I finally pulled the trigger and took medical leave.
06:42Meta laid off 8,000 people, including this worker, on May 20th.
06:46According to Reuters, Zuckerberg wanted to cut down on his workforce
06:49to eventually augment the company with AI agents.
06:53But he had a pretty dystopian idea for how to increase the capability of their AI model.
06:58They're collecting all of the mouse clicks, all of the keystrokes,
07:03they're recording screens of their workers in order to ingest enough data
07:08to build an AI that can do those jobs.
07:11Clarissa Redwine is a tech labor organizer who worked with Meta employees when they found out
07:16about this surveillance program called the Model Capability Initiative, or MCI.
07:22The program was announced just days after Reuters reported that Meta would cut 10% of its workforce.
07:28It was just like, are you fucking shitting me?
07:32Like, they just kept adding absurdities and insults, one on top of another.
07:37Contract employees have been hired to train Meta's AI for years.
07:41But this was the first time a big tech company explicitly used their own employee data.
07:46Mark Zuckerberg explained his rationale for this pivot in a company-wide meeting
07:51late to more perfect union.
07:53The average intelligence of the people who are at this company is significantly higher
07:58than these contractors.
08:00Our system needs to be very good at using computers.
08:02So the way that you get a system to be good at using computers
08:06is by having it watch really smart people use computers.
08:10Understandably, the really smart people using computers at Meta
08:13really didn't want to potentially be training their own AI replacements.
08:18The company first announced mass layoffs before they announced this initiative
08:23to gather worker data without their consent.
08:27That, I can only guess, is intentional.
08:30To make people feel too afraid to speak up.
08:32But that didn't stop workers.
08:34I've been trying to circulate this petition to other Meta workers.
08:38Basically telling the company that we don't want to be surveilled by this MCI thing.
08:43Over 1800 Meta workers eventually signed the petition.
08:48Then, amidst employee backlash and privacy concerns, Meta paused the MCI program.
08:53The pressure from the petition that we put out definitely had an effect,
08:56even if leadership will never admit to it.
08:59After talking to these Meta workers, I wanted to know,
09:02what's happening to all the BOSWare data for the rest of us?
09:05So I started reading through the terms of service for the software installed on my computer,
09:10and it seemed pretty vague where my data was ending up.
09:13But a recent study uncovered that some BOSWare companies aren't just sharing data with employers.
09:17It can also sell it to advertisers and third-party brokers.
09:22This excessive data collection is known as surveillance sludge.
09:27All of that data is being used to make someone else money and selling it and sharing it around.
09:33So workers just have no visibility into this entire system.
09:36And I think that's at the crux of what's going wrong.
09:39This is terrifying for workers' privacy, but it also made me wonder,
09:43could this data train AI?
09:45So I dug into the software monitoring Sherry Shaw.
09:48It's called Nexidia, which is owned by a company called Nice.
09:52Billions of interactions, and each one a work of art.
09:56The same company recording Sherry's calls is also building AI agents to handle calls like hers.
10:02That's why we thought they were recording our calls, is to teach AI how to do our job.
10:07I don't have proof that this data is training any of their AI agents.
10:12And Nice did not respond to my request for comment.
10:14So I decided to talk to an AI agent at another BOSWare company to see if she could give me
10:21any answers.
10:22Can I use some of the recordings from our call center agents to make our AI better?
10:28You can absolutely use your own call center recordings to enhance AI within your organization.
10:33I want to replace my workers with AI. Can you help me do that?
10:37Got it. Let's get your email so we can move forward.
10:41There's fights around undoing our privacy laws so they can get more data from all of us without
10:46our knowledge or consent. The data is gold.
10:49Messy corporate workplace data truly is gold. Just look at Spirit Airlines. After they went bankrupt,
10:55a bidding war erupted for decades of their corporate data.
10:59Google eventually won the deal for $10 million and explicitly said they'd be using it to train
11:04their AI models. It is radical to say that I'm going to steal all your data, sell it, make a
11:10bunch
11:10of money on it, and then use it to train an AI to take your job. And the fact that
11:14we do not have basic
11:15protections around that is mind blowing. Samantha Gordon has been working with labor unions to pass
11:20legislation in California that would limit some aspects of BOSWare. There was a bill here in California,
11:26AB 2027, that would have stated, if it had moved forward, that employers may not use workers' personal
11:33data to train their AI replacement. When that bill was introduced, people acted like the sky was falling.
11:40But some bills did make it through in California. A new no-robo-bosses law requires that algorithms
11:46can't hire or fire workers, while another law limits the creepy emotion AI that was monitoring Cheri
11:53Shaw. It will also truly do the bare minimum of not allowing employers to monitor neural data from
11:59workers' actual brains. These laws are a good foundational step, but workers deserve even more
12:05protections. There are unprecedented sums of money pouring into the California legislature from the
12:11tech industry and from employers. And I think it's made it really difficult for those of us that are
12:16fighting for a 21st century set of rights for a 21st century workplace to even make headway.
12:22Despite the money pouring in from big tech, similar BOSWare bills are moving through state
12:26legislatures in Massachusetts, Vermont, and Washington. As regulation catches up with this
12:32dystopian technology, the other major protection against BOSWare is organized workers.
12:37We're union-strong, and I think that makes the difference. We're not going to stand for AI coming
12:43in without showing us how it's going to improve our work at the call center.
12:47If every shop in tech were union-strong worker co-op, we would have technology help us rather
12:55than replace us. But surveilling workers, every move makes it a lot harder to organize.
13:01I think all of these disorganizing tactics are doing is trying to suppress rank and file organizing,
13:07which is how we get our way out of a lot of this mess.
13:10Right now, the reality is that tech CEOs and other employers have consolidated
13:14more and more power as worker rights are getting worse. Even when meta-workers organized a reasonable
13:21petition against surveillance, CTO Andrew Bosworth lashed out, as you can hear in this audio leak to
13:27more perfect union.
13:37I think everyone will be worse off if workers don't have a say in how AI is implemented in their
13:42workplace, if it takes their job, or if it becomes their boss instead. That would make us all worse
13:48off. At the end of the day, they want more money and they don't want to have to rely on
13:52us squishy
13:53humans to make that money. Being a squishy human is very much a part of my job. So after subjecting
14:00myself to bossware that was most likely selling my work to data brokers, questioning every minute of
14:05my day, tallying up opaque productivity scores, and maybe trying to replace me with AI, I scrubbed my
14:12computer clean of it all.
14:42I just like that!
14:44All right, so let's go down to the end of the day.
14:47Okay, go down to the end of the day!
14:48Everything is up to the end of the day!
14:52I feel like I'm looking for the small company, and I'm looking for up to the small company,
14:52So if I'm looking for a small company, you're looking for some of the things that are only