This Saturday (MAY 30), I did something I never thought possible.

I interviewed one of my biggest role models: Parmy Olson. Yep. THE award-winning Bloomberg journalist who has spent the past 10 years interviewing the AI elite, including Sam Altman, Demis Hassabis, Andrew Ng, Noam Shazeer, Shane Legg, and hundreds of insiders at OpenAI, DeepMind, Google, Anthropic, and Microsoft.

Parmy is one of the few people who know the ins and outs of Silicon Valley. She has focused most of her career on highlighting the contradictions among AI leaders and calling out their bad behavior, even before ChatGPT existed and before anti-AI sentiment began to brew. Truly, she is ONE of the OGs paving the way for me (and other AI ethicists) to do our work. Needless to say, Saturday was one of the highlights of my career, and dare I say, my life!

Before I share the MINDBLOWING intel she gave us (how Anthropic is full of BS, how AI leaders used fear-mongering marketing to sell subscriptions, and how the ideology of effective altruism permeates Silicon Valley), I want to take a step back to thank you. This wouldn't be possible without you.

THANK YOU, THANK YOU & THANK YOU! 💖

Exactly one year ago, I started creating content to help people stay safe in the age of AI. I was tired of tech bros owning the AI narrative and sharing hype-driven “opinions” without a drop of critical thought. But NEVER, in a million years, did I think I would ever reach +700k followers across social media and 20 million views a month. I NEVER imagined I would create a responsible AI literacy community and book club (The AI Survival Club) with 400 members from 25+ countries, where we would read one book on responsible AI every month, and I would get the chance to INTERVIEW the authors themselves. 

I NEVER imagined this would be my life.

But it is.

AND if I am here today, doing what I love, it's because of YOU.

THANK YOU! 

(okay, now let's get to the good stuff!)

Three Key Insights From My Conversation with Parmy Olson

Seventy members of my book club joined me on a live Zoom call for a 50-minute conversation with Parmy Olson about her book, Supremacy, our May read, and the current state of AI. Here are three insights I think you need to hear to better understand the industry and cut through the hype. Let's get started!

[Okay, sorry. Now, let's start. Here are the three insights you need to know.]

1. Sam Altman & Demis Hassabis said they wanted to save humanity. They built a better ad machine.

In Supremacy, Olson presents Sam Altman and Demis Hassabis as the main characters in the quest for AGI (Artificial General Intelligence). To talk about the present (who they are and what they are building), she starts her book by giving us intel about their pasts: tracing their childhoods, hopes, dreams, and motivations. Parmy even interviews Sam Altman's school director to better understand who he was. [Did you know Sam was responsible for the opening of the first LGBTQ+ support group in his school? Yep. I learned this reading Supremacy!]

Now, going back to their roots means we, the readers, are faced with the people they used to be—or, better, who they SAID they were. Back to a time when their goal was to build AGI that would SAVE humanity, PROTECT us from harm, and PUSH society FORWARD. And so, a few chapters in, we were left wondering: were Altman and Hassabis ever truly altruistic, or was it all just a façade?

So I asked Parmy.

Her reply?

She cannot be certain, of course, since she cannot read their minds. But she pointed to the clear gap between their WORDS and ACTIONS.

This is what she said:

"What is Demis Hassabis spending the vast majority of his time with?
It is not actually trying to build AGI that's going to unlock the mysteries of the universe and science. It is improving Gemini and making it one of the most successful new products that Google has ever brought to market, fundamentally changing Google search and online advertising, and how we search for products, and how we use the internet.

That's what he's spending all his time doing.

And Sam's motivation was for a long time about power for himself, glory for himself, and wealth, but also power, glory, and wealth for his staff and for the public. He talked about creating an abundance of wealth for many, many more people. But I think in running OpenAI now, heading towards an IPO potentially in months, partnering with Microsoft, all of that has just kind of gone out the window because of the commercial imperative that he now faces."

Okay, Catharina, were they altruistic, or is this all a façade?

Again, we can't know for sure, but what is CLEAR is that what they SAID they would do (save humanity) and what they are DOING NOW (creating AI to power ads and chatbots) are two VERY different things. If they were ever truly good, their utopian ideals were completely drowned out by commercial pressures the moment they aligned with trillion-dollar companies.

Now, WHY does knowing their contradictions matter?

Because the AI utopia they talk about is the DREAM they USE to sell us their AI.

They talk about it in interviews.
They speak about it in keynotes.
They share it at Davos.

But the reality?

It is NEITHER pink NOR filled with ponies (unfortunately).

They are building AI to make money. NOT to save us.

2. Effective altruism is Silicon Valley’s permission slip to never leave the condo.

In Supremacy, Parmy writes extensively about the effective altruism movement and its hold on Silicon Valley.

[In case you are not familiar, effective altruism is the idea that you can do the most good in the world, not necessarily by directly helping people, but by making AS MUCH MONEY as possible and then donating it in the most optimized way. For example, under effective altruism, it would be more optimized for folks to go to business school and become billionaires to donate to different charities than to go to med school to become doctors and help sick people. If this sounds horrible, it's because it is.]

Well, in her book, Parmy explains how the ideology of effective altruism has infiltrated Silicon Valley, leading tech leaders to focus on protecting trillions of people who haven't yet been born rather than the billions of people living on Earth today who are suffering from the consequences of AI.

Again, if we think about it, optimizing for the good means that trillions of people who HAVEN'T been born YET are worth more than billions of people already living today. 

During our call, I was curious to understand: Why does effective altruism have such a grip on Silicon Valley specifically, and how did it become so widespread?

In her answer, Parmy said that in tech, there is a tendency to justify the money people make as good for humanity, so they never have to think about the “unintended consequences of what they are building.”

She compared working at these companies to being in a cult, where everyone talks about how great the company is and nobody wants to discuss the potential side effects of their algorithms.

"This is an idea that appeals to people who are gonna make a lot of money. That if I'm gonna make a lot of money, I can be much more effective donating that and buying thousands of nets to help prevent malaria in Africa, rather than join Doctors Without Borders, train to be a doctor and fly out to Africa or some other struggling part of the world, and try and help people face to face. I can be more effective if I just stay in my air-conditioned condo in Palo Alto and with my MacBook, and get a six-figure salary, or even a seven-figure salary for working for Alphabet, or OpenAI or Anthropic, and then donating that money. I'll be much more effective doing that. That's why it's called effective altruism."

Why should YOU know this?

You need to understand that effective altruism is rooted in everything Silicon Valley builds, because only then will you START to see WHY they are NOT focused on protecting the HUMANITY OF TODAY, but on the humanity of the FUTURE. 

It is also important because this way of thinking frames morality as a math problem rather than a political or ideological one. In a way, it removes humanity from the equation, because “it’s just math.” 

Lastly, effective altruism quantifies charity (or good) in a way that feels rational, which, when you think about it, is the perfect excuse to keep building things they have no idea how to control.

[Of course, effective altruism is NOT necessarily what they truly believe in, but an ideological façade they use to make it sound logical and to make what their actions do palatable to the audience. Now you know their trick.]

3. Anthropic is NOT the good guy, and they know exactly what they are doing.

Of course, I couldn’t interview Parmy without asking what she thinks about Anthropic. Since her book was published, Anthropic has become one of the most powerful players in AI and has spent a LOT of money positioning itself as the ethical alternative. Literally, I get dozens of questions EVERY SINGLE DAY asking whether I think Anthropic (and Claude) are the good options in the market. 

Of course, I have my own opinions about them (no, I don't believe they are the good guys), but I asked Parmy directly. If anyone knows about this, it's her.

She did not hesitate.

Her answer is below, for paid members. 👀👇

A quick note: the full interview, including why Demis wanted to build AGI to find God, what Jared Kaplan actually told Parmy when she asked why they left OpenAI, and the decision that put Anthropic on the verge of profit while OpenAI burns through cash, is behind a paywall. I know that's not for everyone, and I respect that. But paid memberships are what keep this work independent. If you've ever wondered how I can say the things I say about Anthropic, OpenAI, and the rest of them without losing access or advertiser money, this is how. Your support keeps me independent. Thank you! [The full 50-minute Q&A VIDEO and FULL TRANSCRIPT are also below the paywall]

Read more

My name is Catharina Doria, and I am an AI ethicist with over 580,000 followers, teaching millions how to stay safe in the age of AI. And guess what? On a warm, sunny day, an AI from Meta deactivated one of my Instagram accounts.

The stated reason?

My account was allegedly in violation of “community standards.”

Yep.

(An AI ethicist going against community standards… makes it make sense!)

Since I haven't gotten any proper answers from anyone (Meta, their safety team, or an AI chatbot), I decided to share my story: the story of how an AI might destroy my life.

Let's get going, shall we?

It all started when I adopted my dog, Miss Petunia (or, as I call her, Miss P!).

Miss P. is a 2.5 kg Chihuahua I adopted one month ago.

Until then, she had spent her entire life locked in a cage, used as a breeding dog on a puppy farm. As expected, living in precarious conditions and being hurt every day left her deeply traumatized. 

On her first day home, she would pretend to be dead, squish herself on the floor, and urinate out of fear. Even when I fed her, she would overfill her little mouth with food, like a little hamster or a squirrel. She would run to a corner, spit out her food, and eat in peace, afraid I would take it away from her. Truly, Miss P. was afraid of her own little shadow.

Granted, adopting her made me really angry with the horrible humans who hurt her AND with the MANY OTHER HUMANS who hurt other dogs just like her. With all that bottled-up anger about animal cruelty, I started sharing her story with my 580K followers.

To my surprise, I received the kindest response ever. Everyone loved Miss P. and wanted to learn more about her story. I was then kindly asked by thousands of my followers to create an Instagram account just for her so everyone could see her progress and her adaptation to her new, very spoiled life.

Creating an account for Miss P. sounded like a wonderful idea.

What could go wrong, right? 

Well, everything.

The moment I pressed “Create Account,” her account was deactivated.

Why, you might ask?

According to the email I received, my new profile was not adhering to their “community standards on account integrity.”

Which ones?

Well, I have NO idea because they didn't tell me.

Mind you: I did not have time to write a bio or add a profile picture. I wrote her username (@misspetuniathechi) and shared some of MY personal details, and that was it. On my end, I truly cannot understand what could have caused my account to be deactivated.

And before you say dogs can't have Instagram accounts, creating one for a dog does NOT violate the community guidelines. In fact, there are MANY dog influencers on Instagram with millions of followers. 

The only thing I could think of is that an “AI” tasked with finding spammy accounts flagged my dog’s profile.

But of course, that's speculation, since NO ONE [no email, no AI, no chat] could tell me what I did wrong.

So I tried appealing.

Again, something was clearly wrong, and it was obvious that an AI had made a mistake. I pressed all the buttons: I shared my face, my name, and my passport. I wrote a blurb explaining the entire situation, how this was all a big mistake, hoping someone would see how unfair it was. 

The result? 

A big sack of nothing. 

A few minutes later, I received an email saying that, upon “closer inspection”, I was REALLY violating the community standards. 

“What does that mean, Catha!?”

I have no clue.

From my side, I can tell you AGAIN:

I checked everything there was to check. There’s no community guideline stating that a dog cannot have an Instagram account. I didn't write the bio. I didn't have time to add a picture. I couldn’t (and still can't) understand what I did wrong. So even if I did something wrong (by mistake), no one could tell me what happened. So I did what any normal human being would do: cry it out in pure frustration.

And then things got exponentially worse.

⭐️ THE DOMINO EFFECT

Maybe you don't know this, but if ONE of your Instagram accounts is deactivated, ALL OF YOUR OTHER ACCOUNTS in your account center are deactivated as well. Plus, you’re NOT allowed to create any new Instagram accounts ever. You get banned from everything! 

And that happened to me.

Three weeks later, my book club account, where I teach thousands of people critical AI literacy (The AI Survival Club) and where I make most of my living, was banned.

The stated reason? 

I had ONE “violating” account in my account center.

Yep.

Miss Petunia's account.

And this was the moment I realized I was in big trouble.

Why?

Because now I have TWO banned accounts in my account center, which means my MAIN account with 580,000 followers could and can be banned at any moment, too, for the same reason: a dog account that was deactivated the second it was created. And now a SECOND banned account, a book club account, was banned because a dog's account was banned.

Talk about a catch-22! 🙃

Long story short, what was supposed to be a wonderful story about an adopted dog became a true nightmare.

The more I think about it, the more it feels like a sick joke from the Universe. I’m an AI ethicist who talks about the dangers of opaque AI, and I got banned because a random AI decided that my dog’s account violated community guidelines, with no transparency. 

Of course, if I am trying to see the glass as half full, this only underscores the importance of my work (and that of other wonderful AI ethicists pushing for fairer AI systems) and shows how AI systems TODAY are making life-altering decisions about people, with no transparency, no human in the loop, and no real appeals process. It sort of beautifully proves the point I make every single day as an AI ethicist.

AI is making life-altering decisions about our LIVES with no accountability, responsibility, or transparency.

But it's not because this is proving me right that I am happy (if you know me, you know I love to be right, but not at the cost of losing my livelihood to a non-transparent AI)

If I am being completely honest with you, I am angry.

Nay, I am pissed.

And what bugs me the MOST is that:

1) I’m an AI ethicist,
2) I talk about AI harm every single day,
3) I have THOUSANDS of followers, MILLIONS OF VIEWS,
4) I have media coverage (thank you, BBC, for covering this case!)

AND I STILL GOT NO PROPER REPLY FROM ANYONE. 

If an AI ethicist with almost 600K followers and press coverage on BBC for a banned account STILL cannot get a clear answer from Meta, what hope does anyone else have?

This opens a bigger conversation.

⭐️ BIG TECH HOLDS TOO MUCH POWER

How can corporations hoard so much power over people’s lives? How can an artificial intelligence system make life-altering decisions about my life—deciding whether I’ll have a job next week and whether I’ll be able to pay my rent—with ZERO transparency, zero accountability, zero anything?

How can AI decide, do, and be done?

I have savings, thank God. 

But imagine the people who don’t. Those who rely on Meta or other platforms for monetization.

What happens to them? 

As an AI ethicist, I find it absurd that a content creator who posts +15 videos a month and treats it as a full-time job can, in a way, be “fired” for no reason (or, if there is a reason, not know which one).

Right now, I worry every day that my account will be destroyed. I am having emergency sessions with a psychologist and taking medication to help me relax. This has completely disrupted my workflow. 

My life became a living hell.

And since my life is now a living hell, I want to make this a learning experience for ALL of us.

⭐️ THE LESSON I’VE LEARNED (AND I HOPE YOU LEARN IT TOO!)

The biggest lesson I've learned from this is that I (and we) can't rely on them for anything. Social media platforms, in general, have repeatedly shown how little respect they have for their content creators and the people working full-time on their platforms.

That is why I am here today to say we shouldn't put all our eggs in one basket. If you create content and have an audience there, please, I urge you to build an audience on a platform that lets you build a mailing list. That is the ONLY way to “own” your audience and ensure THEY can't take them away from you. 

Please. Build. Your. Email. List.

And I know, now you might be thinking: 

“But Catharina, you are an AI ethicist… You didn’t think of this before?”

Of course, I knew this in theory; I just somehow never thought it would happen to me. I know it sounds ridiculous to say this out loud (or in writing), but I somehow thought that if I played by the rules of the game, the game wouldn’t screw me over.

I was clearly wrong.

Lesson learned.

⭐️ LEARNING THROUGH LOVE, PAIN, AND OBSERVATION

You see, my dad (may God rest his soul) always told me that people learn in three ways: through pain, love, or observation.

• Love is when someone tells you something kindly, and you accept their advice.

• Pain is when you don't listen, get hurt, and learn your lesson.

• Observation is when you look around, see what happened to others, and learn from it.

For example: one day, your mom tells you NOT to touch the hot stove because you could get hurt. You can…

1) accept her advice because she's your mom (love)

2) put your hand on the hot stove, burn yourself, and never do it again (pain), or

3) see your sister put her hand on the hot stove, burn herself, and decide not to do it yourself because you saw how badly she was hurt (observation)

If it's not clear, I am your sister who burned her hands, and now you get to learn from my burn. I had to learn the painful way that I cannot trust social media platforms. Now is your chance to learn either through love (if you trust me) or by observing, rather than repeating my mistake. 

Please don’t procrastinate on this decision and build your mailing list.

⭐️ MY SUBSTACK JOURNEY STARTS HERE

Okay, what does that mean in practice? Well, from now on, you will see me here on Substack. I will put the same effort here as on my Instagram. If they censor me there and remove me from their platform, I will be here, standing and talking about ethical AI and teaching people to protect themselves from this technology (and from the people behind it).

I hope you will stay with me here, too.

Today, if you are receiving this email, it is because you filled out a form to join my mailing list or saw my stories and chose to follow me here as well. If you haven't subscribed yet and have read this far, now would be a good time to subscribe. I hope I can be worthy of your trust.

⭐️ ALL IN ALL: WE KEEP FIGHTING!

As a final note, I want to thank BBC News for covering this story; journalist Luiz Fernando Toledo for pushing Meta for answers; Caio Vieira Machado for his expert advice; and my digital rights lawyer, Bruna Oliveira, for helping me navigate the right legal channels.

I am still here, and I am not going anywhere without a fight.

And to every creator, small business owner, journalist, activist, or ordinary user who has had their account wrongfully closed by an algorithm with no recourse: I see you. I stand with you.

Companies need to be held accountable for their AI shenanigans ASAP.

I truly hope policymakers, lawmakers, grassroots organizations, and everyone can work together to ensure these companies are held accountable and liable for their actions. 

Until laws change, we, the public, have to keep fighting. And rest assured, I will continue to fight. 😉

⭐️ PLEASE SHARE THIS TO RAISE AWARENESS

If this post resonates with you, please share it.

The more people who see this, the harder it is for Meta (and other social companies making life-altering decisions about our lives with no transparency) to keep ignoring it.

I will see you again on Friday with five responsible AI papers I highly recommend you read if you want to stay protected from AI.

If you're interested in that, don't forget to subscribe. 

Thank you for reading.

Many hugs, 
Catharina Doria

AI Ethicist | Digital Rights Advocate | Content Creator

Making AI make sense: with ethics, clarity, and a little rebellion.

[Also, to be clear: this is all very real. I got a message from a follower (or maybe a hater) asking whether my “deactivation bit” worked to boost engagement. This is not a bit. This is very real, with very real consequences, and I have a very real digital rights lawyer involved in the case to get my accounts back.]