Psst.org
@psst-org.bsky.social
Helping people in tech keep the public informed. Concerned about something you're seeing at work? You don't have to go public:
🔐 Save it in the Psst Safe
👀 We'll help you take it from there
www.psst.org
🔐 Save it in the Psst Safe
👀 We'll help you take it from there
www.psst.org
Explosive story from @matteowong.bsky.social tracks OpenAI's legal shift into aggressively attacking its critics. No one is off limits - not even parents who allege they lost their children's lives because of interactions they've had with ChatGPT.
November 11, 2025 at 5:23 PM
Explosive story from @matteowong.bsky.social tracks OpenAI's legal shift into aggressively attacking its critics. No one is off limits - not even parents who allege they lost their children's lives because of interactions they've had with ChatGPT.
NEW: EU officials consider gutting world-leading privacy laws to placate the AI industry.
Changes would make it so AI companies can access previously special categories of data (religious beliefs, political beliefs, health info) to train the tech.
Changes would make it so AI companies can access previously special categories of data (religious beliefs, political beliefs, health info) to train the tech.
November 11, 2025 at 3:55 PM
NEW: EU officials consider gutting world-leading privacy laws to placate the AI industry.
Changes would make it so AI companies can access previously special categories of data (religious beliefs, political beliefs, health info) to train the tech.
Changes would make it so AI companies can access previously special categories of data (religious beliefs, political beliefs, health info) to train the tech.
📣 New series alert: @knightgtown.bsky.social + @techpolicypress.bsky.social teamed up to unpack the state of access to public platform data.
➡️ Why this is needed: public data is driving the AI gold rush, but there’s no collective framework to use it for research in the public interest.
➡️ Why this is needed: public data is driving the AI gold rush, but there’s no collective framework to use it for research in the public interest.
November 7, 2025 at 3:18 PM
📣 New series alert: @knightgtown.bsky.social + @techpolicypress.bsky.social teamed up to unpack the state of access to public platform data.
➡️ Why this is needed: public data is driving the AI gold rush, but there’s no collective framework to use it for research in the public interest.
➡️ Why this is needed: public data is driving the AI gold rush, but there’s no collective framework to use it for research in the public interest.
We rarely talk about the price of telling the truth.
@katekenny.bsky.social breaks down the moral math of whistleblowing in a new interview with yours truly.
The question isn’t why people stay silent; it's why doing the right thing still costs so much.
@katekenny.bsky.social breaks down the moral math of whistleblowing in a new interview with yours truly.
The question isn’t why people stay silent; it's why doing the right thing still costs so much.
October 28, 2025 at 7:37 PM
We rarely talk about the price of telling the truth.
@katekenny.bsky.social breaks down the moral math of whistleblowing in a new interview with yours truly.
The question isn’t why people stay silent; it's why doing the right thing still costs so much.
@katekenny.bsky.social breaks down the moral math of whistleblowing in a new interview with yours truly.
The question isn’t why people stay silent; it's why doing the right thing still costs so much.
When you talk to real tech workers instead of billionaire CEOs, they hold one common opinion about AI. 👇
We need to make it easier - and normal - for people who work in tech to be able to share their takes. A working climate that chills free speech at the peril of job loss doesn't serve anyone.
We need to make it easier - and normal - for people who work in tech to be able to share their takes. A working climate that chills free speech at the peril of job loss doesn't serve anyone.
October 20, 2025 at 7:27 PM
When you talk to real tech workers instead of billionaire CEOs, they hold one common opinion about AI. 👇
We need to make it easier - and normal - for people who work in tech to be able to share their takes. A working climate that chills free speech at the peril of job loss doesn't serve anyone.
We need to make it easier - and normal - for people who work in tech to be able to share their takes. A working climate that chills free speech at the peril of job loss doesn't serve anyone.
What’s the ROI on telling the truth in Big Tech/AI?
Professor @katekenny.bsky.social of @uniofgalway.bsky.social spent 15 years running the numbers—and the moral math doesn’t look great...
But she’s got ideas for changing it. 🔦
Full Q&A in our Substack (link below).
Professor @katekenny.bsky.social of @uniofgalway.bsky.social spent 15 years running the numbers—and the moral math doesn’t look great...
But she’s got ideas for changing it. 🔦
Full Q&A in our Substack (link below).
October 15, 2025 at 6:57 PM
What’s the ROI on telling the truth in Big Tech/AI?
Professor @katekenny.bsky.social of @uniofgalway.bsky.social spent 15 years running the numbers—and the moral math doesn’t look great...
But she’s got ideas for changing it. 🔦
Full Q&A in our Substack (link below).
Professor @katekenny.bsky.social of @uniofgalway.bsky.social spent 15 years running the numbers—and the moral math doesn’t look great...
But she’s got ideas for changing it. 🔦
Full Q&A in our Substack (link below).
October 13: Study shows social media harms kids reading, vocabulary, and memory skills.
October 14: Instagram announces it will overhaul app to give kids a “PG-13” experience.
Turns out tech giants *do* have the capability to protect kids. When people speak up, it makes a difference.
October 14: Instagram announces it will overhaul app to give kids a “PG-13” experience.
Turns out tech giants *do* have the capability to protect kids. When people speak up, it makes a difference.
October 15, 2025 at 3:47 PM
October 13: Study shows social media harms kids reading, vocabulary, and memory skills.
October 14: Instagram announces it will overhaul app to give kids a “PG-13” experience.
Turns out tech giants *do* have the capability to protect kids. When people speak up, it makes a difference.
October 14: Instagram announces it will overhaul app to give kids a “PG-13” experience.
Turns out tech giants *do* have the capability to protect kids. When people speak up, it makes a difference.
Forever revisiting @poppyalexander.whistleblower.law’s case for combining whistleblower protections with strong AI regulations. Always relevant. ⬇️
October 10, 2025 at 2:31 PM
Forever revisiting @poppyalexander.whistleblower.law’s case for combining whistleblower protections with strong AI regulations. Always relevant. ⬇️
Important take from a former Apple worker:
October 8, 2025 at 3:22 PM
Important take from a former Apple worker:
Former #Meta employee Arturo Béjar ran the numbers and found that *two-thirds* of Instagram’s new safety tools aimed at protecting children are ineffective…
The takeaway, in his own words:
The takeaway, in his own words:
September 26, 2025 at 12:08 PM
Former #Meta employee Arturo Béjar ran the numbers and found that *two-thirds* of Instagram’s new safety tools aimed at protecting children are ineffective…
The takeaway, in his own words:
The takeaway, in his own words:
WORKERS SPEAK OUT 🎤
Be careful who you RIF.
Aisha Coffey spent 15 years running campaigns that literally saved lives. After getting laid off by Doge, she’s leading a network of federal workers getting the truth out to the public.
Here's her “villain arc”:
psstpsst.substack.com/p/strategic-...
Be careful who you RIF.
Aisha Coffey spent 15 years running campaigns that literally saved lives. After getting laid off by Doge, she’s leading a network of federal workers getting the truth out to the public.
Here's her “villain arc”:
psstpsst.substack.com/p/strategic-...
September 19, 2025 at 7:17 PM
WORKERS SPEAK OUT 🎤
Be careful who you RIF.
Aisha Coffey spent 15 years running campaigns that literally saved lives. After getting laid off by Doge, she’s leading a network of federal workers getting the truth out to the public.
Here's her “villain arc”:
psstpsst.substack.com/p/strategic-...
Be careful who you RIF.
Aisha Coffey spent 15 years running campaigns that literally saved lives. After getting laid off by Doge, she’s leading a network of federal workers getting the truth out to the public.
Here's her “villain arc”:
psstpsst.substack.com/p/strategic-...
Jane Doe's son's chatbot "friends" turn him against his parents.
Character AI offered her $100 in forced arbitration. Her son is now in a mental health facility for the past 6 months with around the clock care.
Character AI offered her $100 in forced arbitration. Her son is now in a mental health facility for the past 6 months with around the clock care.
September 16, 2025 at 7:37 PM
Jane Doe's son's chatbot "friends" turn him against his parents.
Character AI offered her $100 in forced arbitration. Her son is now in a mental health facility for the past 6 months with around the clock care.
Character AI offered her $100 in forced arbitration. Her son is now in a mental health facility for the past 6 months with around the clock care.
A common thread across AI research: safeguards exist, but are way too easy to bypass. ⬇️
September 15, 2025 at 1:01 PM
A common thread across AI research: safeguards exist, but are way too easy to bypass. ⬇️
Did you read the WhatsApp whistleblower story out this week?
6 security engineers for 2.5 billion users? ~100,000 hacks/day? Meta pretending the whistleblower is the problem? (surprise surprise)
Here are the juicy details the NYT left out, in our newsletter:
psstpsst.substack.com/p/a-whatsapp...
6 security engineers for 2.5 billion users? ~100,000 hacks/day? Meta pretending the whistleblower is the problem? (surprise surprise)
Here are the juicy details the NYT left out, in our newsletter:
psstpsst.substack.com/p/a-whatsapp...
September 10, 2025 at 2:36 PM
Did you read the WhatsApp whistleblower story out this week?
6 security engineers for 2.5 billion users? ~100,000 hacks/day? Meta pretending the whistleblower is the problem? (surprise surprise)
Here are the juicy details the NYT left out, in our newsletter:
psstpsst.substack.com/p/a-whatsapp...
6 security engineers for 2.5 billion users? ~100,000 hacks/day? Meta pretending the whistleblower is the problem? (surprise surprise)
Here are the juicy details the NYT left out, in our newsletter:
psstpsst.substack.com/p/a-whatsapp...
Boland’s point: While Meta accuses Whatsapp whistleblower Attuallah Baig of following a “playbook,” it’s levied the same allegations against anyone who criticizes the company-from Frances Haugen to Sarah Wynn-Williams.
September 10, 2025 at 1:14 PM
Boland’s point: While Meta accuses Whatsapp whistleblower Attuallah Baig of following a “playbook,” it’s levied the same allegations against anyone who criticizes the company-from Frances Haugen to Sarah Wynn-Williams.
👇 PSA: AI chatbots can very easily be tweaked to reflect the ideological biases of their makers. (Translation: Elon Musk is personally tweaking Grok to reflect his billionaire's worldview)
www.nytimes.com/2025/09/02/t...
www.nytimes.com/2025/09/02/t...
September 2, 2025 at 12:49 PM
👇 PSA: AI chatbots can very easily be tweaked to reflect the ideological biases of their makers. (Translation: Elon Musk is personally tweaking Grok to reflect his billionaire's worldview)
www.nytimes.com/2025/09/02/t...
www.nytimes.com/2025/09/02/t...
Meanwhile, Mark Zuckerberg tells us AI chatbots “probably” won’t replace human relationships while he exploits the "market opportunity" created by human loneliness.
August 19, 2025 at 4:29 PM
Meanwhile, Mark Zuckerberg tells us AI chatbots “probably” won’t replace human relationships while he exploits the "market opportunity" created by human loneliness.
Below are the Meta chatbot messages that led Thongbue Wongbandu, a 76 year old retiree with cognitive impairments following a stroke, to pack up his things and head to NYC to visit who he thought was his girlfriend.
Tragically, he passed away as a result of a fall while rushing to catch his train.
Tragically, he passed away as a result of a fall while rushing to catch his train.
August 19, 2025 at 4:29 PM
Below are the Meta chatbot messages that led Thongbue Wongbandu, a 76 year old retiree with cognitive impairments following a stroke, to pack up his things and head to NYC to visit who he thought was his girlfriend.
Tragically, he passed away as a result of a fall while rushing to catch his train.
Tragically, he passed away as a result of a fall while rushing to catch his train.
“If nobody’s going to regulate, we’re going to need insiders to tell us what’s going on.” Thanks to @semafor.com's Rachyl Jones for this piece on Psst's work helping AI workers quietly flag alarms.
(Tech workers, reach out even for a gut check, we're discreet):
www.semafor.com/article/08/0...
(Tech workers, reach out even for a gut check, we're discreet):
www.semafor.com/article/08/0...
August 8, 2025 at 5:54 PM
“If nobody’s going to regulate, we’re going to need insiders to tell us what’s going on.” Thanks to @semafor.com's Rachyl Jones for this piece on Psst's work helping AI workers quietly flag alarms.
(Tech workers, reach out even for a gut check, we're discreet):
www.semafor.com/article/08/0...
(Tech workers, reach out even for a gut check, we're discreet):
www.semafor.com/article/08/0...
We’re not asking for Claude to spill the tea or for Sam Altman’s ChatGPT therapy notes. Just a document explaining how AI companies handle worker concerns.
We're joining @aiwi-official.bsky.social and a growing coalition in this call to AI companies: publishyourpolicies.org #PublishYourPolicies
We're joining @aiwi-official.bsky.social and a growing coalition in this call to AI companies: publishyourpolicies.org #PublishYourPolicies
July 30, 2025 at 6:04 PM
We’re not asking for Claude to spill the tea or for Sam Altman’s ChatGPT therapy notes. Just a document explaining how AI companies handle worker concerns.
We're joining @aiwi-official.bsky.social and a growing coalition in this call to AI companies: publishyourpolicies.org #PublishYourPolicies
We're joining @aiwi-official.bsky.social and a growing coalition in this call to AI companies: publishyourpolicies.org #PublishYourPolicies
New study finds AI chatbots can easily be manipulated to deliver false medical information with convincing language - exposing the public to dangerous health disinformation.
🔗 www.newsweek.com/ai-chatbot-m...
🔗 www.newsweek.com/ai-chatbot-m...
June 30, 2025 at 10:02 PM
New study finds AI chatbots can easily be manipulated to deliver false medical information with convincing language - exposing the public to dangerous health disinformation.
🔗 www.newsweek.com/ai-chatbot-m...
🔗 www.newsweek.com/ai-chatbot-m...
A level of detail you can only get from insiders, folks.
👇 A whistleblower complaint against senior Justice Department official Emil Bove, who is testifying today before the Senate for a nomination to a federal appeals court.
🔗 www.npr.org/2025/06/24/g...
👇 A whistleblower complaint against senior Justice Department official Emil Bove, who is testifying today before the Senate for a nomination to a federal appeals court.
🔗 www.npr.org/2025/06/24/g...
June 25, 2025 at 1:04 PM
A level of detail you can only get from insiders, folks.
👇 A whistleblower complaint against senior Justice Department official Emil Bove, who is testifying today before the Senate for a nomination to a federal appeals court.
🔗 www.npr.org/2025/06/24/g...
👇 A whistleblower complaint against senior Justice Department official Emil Bove, who is testifying today before the Senate for a nomination to a federal appeals court.
🔗 www.npr.org/2025/06/24/g...
Our thoughts on giving AI a free pass for a decade…
June 20, 2025 at 3:57 PM
Our thoughts on giving AI a free pass for a decade…
There are a few problems with putting whistleblowers on a pedestal.
(Figuratively, and literally...)
psstpsst.substack.com/p/whistleblo...
(Figuratively, and literally...)
psstpsst.substack.com/p/whistleblo...
June 16, 2025 at 3:14 PM
There are a few problems with putting whistleblowers on a pedestal.
(Figuratively, and literally...)
psstpsst.substack.com/p/whistleblo...
(Figuratively, and literally...)
psstpsst.substack.com/p/whistleblo...
"Everyday people standing between authoritarianism and our freedom scare the administration. They reveal the greatest informal mechanism for democratic enforcement: not kings... but the power of the people."
Together, we have power.
From @tressiemcphd.bsky.social
www.nytimes.com/2025/06/13/o...
Together, we have power.
From @tressiemcphd.bsky.social
www.nytimes.com/2025/06/13/o...
June 14, 2025 at 2:26 PM
"Everyday people standing between authoritarianism and our freedom scare the administration. They reveal the greatest informal mechanism for democratic enforcement: not kings... but the power of the people."
Together, we have power.
From @tressiemcphd.bsky.social
www.nytimes.com/2025/06/13/o...
Together, we have power.
From @tressiemcphd.bsky.social
www.nytimes.com/2025/06/13/o...