Connect with us

Technology

Inside the launch — and future — of ChatGPT

Published

on

Inside the launch — and future — of ChatGPT

As winter descended on San Francisco in late 2022, OpenAI quietly pushed a new service dubbed ChatGPT live with a blog post and a single tweet from CEO Sam Altman. The team labeled it a “low-key research preview” — they had good reason to set expectations low. 

“It couldn’t even do arithmetic,” Liam Fedus, OpenAI’s head of post-training says. It was also prone to hallucinating or making things up, adds Christina Kim, a researcher on the mid-training team.

Ultimately, ChatGPT would become anything but low-key.

While the OpenAI researchers slept, users in Japan flooded ChatGPT’s servers, crashing the site only hours after launch. That was just the beginning.

“The dashboards at that time were just always red,” recalls Kim. The launch coincided with NeurIPS, the world’s premier AI conference, and soon ChatGPT was the only thing anyone there could talk about. ChatGPT’s error page — “ChatGPT is at capacity right now” — would become a familiar sight.

Advertisement

“We had the initial launch meeting in this small room, and it wasn’t like the world just lit on fire all of a sudden,” Fedus says during a recent interview from OpenAI’s headquarters. “We’re like, ‘Okay, cool. I guess it’s out there now.’ But it was the next day when we realized — oh, wait, this is big.”

“The dashboards at that time were just always red.”

Two years later, ChatGPT still hasn’t cracked advanced arithmetic or become factually reliable. It hasn’t mattered. The chatbot has evolved from a prototype to a $4 billion revenue engine with 300 million weekly active users. It has shaken the foundations of the tech industry, even as OpenAI loses money (and cofounders) hand over fist while competitors like Anthropic threaten its lead.

Whether used as praise or pejorative, “ChatGPT” has become almost synonymous with generative AI. Over a series of recent video calls, I sat down with Fedus, Kim, ChatGPT head of product Nick Turley, and ChatGPT engineering lead Sulman Choudhry to talk about ChatGPT’s origins and where it’s going next.

A “weird” name and a scrappy start

Advertisement

ChatGPT was effectively born in December 2021 with an OpenAI project dubbed WebGPT: an AI tool that could search the internet and write answers. The team took inspiration from WebGPT’s conversational interface and began plugging a similar interface into GPT-3.5, a successor to the GPT-3 text model released in 2020. They gave it the clunky name “Chat with GPT-3.5” until, in what Turley recalls as a split-second decision, they simplified it to ChatGPT. 

The name could have been the even more straightforward “Chat,” and in retrospect, he thinks perhaps it should have been. “The entire world got used to this odd, weird name, we’re probably stuck with it. But obviously, knowing what I know now, I wish we picked a slightly easier to pronounce name,” he says. (It was recently revealed that OpenAI purchased the domain chat.com for more than $10 million of cash and stock in mid-2023.)

As the team discovered the model’s obvious limitations, they debated whether to narrow its focus by launching a tool for help with meetings, writing, or coding. But OpenAI cofounder John Schulman (who has since left for Anthropic) advocated for keeping the focus broad.

The team describes it as a risky bet at the time; chatbots were viewed as an unremarkable backwater of machine learning, they thought, with no successful precedents. Adding to their concerns, Facebook’s Galactica AI bot had just spectacularly flamed out and been pulled offline after generating false research.

The team grappled with timing. GPT-4 was already in development with advanced features like Code Interpreter and web browsing, so it would make sense to wait to release ChatGPT atop the more capable model. Kim and Fedus also recall people wanting to wait and launch something more polished, especially after seeing other companies’ undercooked bots fail.

Advertisement

Despite early concerns about chatbots being a dead end, The New York Times has reported that other team members worried competitors would beat OpenAI to market with a fresh wave of bots. The deciding vote was Schulman, Fedus and Kim say. He pushed for an early release, alongside Altman, both believing it was important to get AI into peoples’ hands quickly.

OpenAI had demoed a chatbot at Microsoft Build earlier that year and generated virtually no buzz. On top of that, many of ChatGPT’s early users didn’t seem to be actually using it that much. The team shared their prototype with about 50 friends and family members. Turley “personally emailed every single one of them” every day to check in. While Fedus couldn’t recall exact figures, he recalls that about 10 percent of that early test group used it every day.

Image: Cath Virginia / The Verge, Getty Images

Later, the team would see this as an indication they’d created something with potential staying power.

“We had two friends who basically were on it from the start of their work day — and they were founders,” Kim recalls. “They were on it basically for 12 to 16 hours a day, just talking to it all day.” With just two weeks before the end of November, Schulman made the final call: OpenAI would launch ChatGPT on the last day of that month.

Advertisement

The team canceled their Thanksgiving plans and began a two-week sprint to public release. Much of the system was built at this point, Kim says, but its security vulnerabilities were untested. So they focused heavily on red teaming, or stress testing the system for potential safety problems. 

“If I had known it was going to be a big deal, I would certainly not want to ship it right before a winter holiday week before we were all going to go home,” Turley says. “I remember working very hard, but I also remember thinking, ‘Okay, let’s get this thing out, and then we’ll come back after the holiday to look at the learnings, to see what people want out of an AI assistant.’”

In an internal Slack poll, OpenAI employees guessed how many users they would get. Most predictions ranged from a mere 10,000 to 50,000. When someone suggested it might reach a million users, others jumped in to say that was wildly optimistic.

On launch day, they realized they’d all been incredibly wrong.

After Japan crashed their servers, and red dashboards and error messages abounded, the team was anxiously picking up the pieces and refreshing Twitter to gauge public reaction, Kim says. They believed the reaction to ChatGPT could only go one of two ways: total indifference or active contempt. They worried people might discover problematic ways to use it (like attempting to jailbreak it), and the uncertainty of how the public would receive their creation kept them in a state of nervous anticipation.

Advertisement

The launch was met with mixed emotions. ChatGPT quickly started facing criticism over accuracy issues and bias. Many schools ran to immediately ban it over cheating concerns. Some users on Reddit likened it to the early days of Google (and were shocked it was free). For its part, Google dubbed the chatbot a “code red” threat.

OpenAI would wind up surpassing its most ambitious 1-million-user target within five days of launch. Two months after its debut, ChatGPT garnered more than 30 million users.

When someone suggested it might reach a million users, others jumped in to say that was wildly optimistic.

Within weeks of ChatGPT’s November 30th launch, the team started rolling out updates incorporating user feedback (like its tendency to give overly verbose answers). The initial chaos had settled, user numbers were still climbing, and the team had a sobering realization: if they wanted to keep this momentum, things would have to change. The small group that launched a “low-key research preview” — a term that would become a running joke at OpenAI — would need to get a lot bigger.

Over the coming months and years, ChatGPT’s team would grow enormously and shift priorities — sometimes to the chagrin of many early staffers. Top researcher Jan Leike, who played a crucial role in refining ChatGPT’s conversational abilities and ensuring its outputs aligned with user expectations, quit this year to join Anthropic after claiming that “safety culture and processes have taken a backseat to shiny products” at OpenAI.

Advertisement

These days, OpenAI is focused on figuring out what the future of ChatGPT looks like.

“I’d be very surprised if a year from now this thing still looks like a chatbot,” Turley says, adding that current chat-based interactions would soon feel as outdated as ’90s instant messaging. “We’ve gotten pretty sidetracked by just making the chatbot great, but really, it’s not what we meant to build. We meant to build something much more useful than that.”

Increasingly powerful and expensive 

I talk with Turley over a video call as he sits in a vast conference room in OpenAI’s San Francisco headquarters that epitomizes the company’s transformation. The office is all sweeping curves and polished minimalism, a far cry from its original office that was often described as a drab, historic warehouse.

With roughly 2,000 employees, OpenAI has evolved from a scrappy research lab into a $150 billion tech powerhouse. The team is spread across numerous projects, including building underlying foundation models and developing non-text tools like the video generator, Sora. ChatGPT is still OpenAI’s highest-profile product by far. Its popularity has come with a lot of headaches. 

Advertisement

“I’d be very surprised if a year from now this thing still looks like a chatbot”

ChatGPT still spins elaborate lies with unwavering confidence, but now they’re being cited in court filings and political discourse. It has allowed for an impressive amount of experimentation and creativity, but some of its most distinctive use cases turned out to be spam, scams, and AI-written college term papers.

While some publications (include The Verge’s parent company, Vox Media) are choosing to partner with OpenAI, others like The New York Times are opting to sue it for copyright infringement. And OpenAI is burning through cash at a staggering rate to keep the lights on.

Turley acknowledges that ChatGPT’s hallucinations are still a problem. “Our early adopters were very comfortable with the limitations of ChatGPT,” he says. “It’s okay that you’re going to double check what it said. You’re going to know how to prompt around it. But the vast majority of the world, they’re not engineers, and they shouldn’t have to be. They should just use this thing and rely on it like any other tool, and we’re not there yet.”

Accuracy is one of the ChatGPT team’s three focus areas for 2025. The others are speed and presentation (i.e., aesthetics).

Advertisement

“I think we have a long way to go in making ChatGPT more accurate and better at citing its sources and iterating on the quality of this product,” Turley says.

OpenAI is also still figuring out how to monetize ChatGPT. Despite deploying increasingly powerful and costly AI models, the company has maintained a limited free tier and a $20 monthly ChatGPT Plus service since February 2023.

When I ask Turley about rumors of a future $2,000 subscription, or if advertising will be baked into ChatGPT, he says there is “no current plan to raise prices.” As for ads: “We don’t care about how much time you spend on ChatGPT.” 

“They should just use this thing and rely on it like any other tool, and we’re not there yet.”

“I’m really proud of the fact that we have incentives that are incredibly aligned with our users,” he says. Those who “use our product a lot pay us money, which is a very, very, upfront and direct transaction. I’m proud of that. Maybe we’ll have a technology that’s much more expensive to serve and we’re going to have to rethink that model. You gotta remain humble about where the technology is going to go.”

Advertisement

Only days after Turley tells me this, ChatGPT did get a new $200 price tag for a pro tier that includes access to a specialized reasoning model. Its main $20 Plus tier is sticking around but it’s clearly not the ceiling for what OpenAI thinks people will pay.

ChatGPT and other OpenAI services require vast amounts of computing power and data storage to keep its services running smoothly. On top of the user base OpenAI has gained through its own products, it’s poised to reach millions of more people through an Apple partnership that integrates ChatGPT with iOS and macOS.

That’s a lot of infrastructure pressure for a relatively young tech company, says ChatGPT engineering lead Sulman Choudhry. “Just keeping it up and running is a very, very big feat,” he says. People love features like ChatGPT’s advanced voice mode. But scaling limitations mean there’s often a significant gap between the the technology’s capabilities and what people can experience. “There’s a very, very big delta there, and that delta is sort of how you scale the technology and how you scale infrastructure.”

Even as OpenAI grapples with these problems, it’s trying to work itself deeper into users’ lives. The company is racing to build agents, or AI tools that can perform complex, multistep tasks autonomously. In the AI world, these are called tasks with a longer “time horizon,” requiring the AI to maintain coherence over a longer period while handling multiple steps. For instance, earlier this year at the company’s Dev Day conference, OpenAI showcased AI agents that could make phone calls to place food orders and make hotel reservations in multiple languages.

For Turley and others, this is where the stakes will get particularly steep. Agents could make AI far more useful by moving what it can do outside the chatbot interface. The shift could also grant these tools an alarming level of access to the rest of your digital life.

Advertisement

“I’m really excited to see where things go in a more agentic direction with AI,” Kim tells me. “Right now, you go to the model with your question but I’m excited to see the model more integrated into your life and doing things proactively, and taking actions on your behalf”

The goal of ChatGPT isn’t to be just a chatbot, says Fedus. As it exists today, ChatGPT is “pretty constrained” by its interface and compute. He says the goal is to create an entity that you can talk to, call, and trust to work for you. Fedus thinks systems like OpenAI’s “reasoning” line of models, which create a trail of checkable steps explaining their logic, could make it more reliable for these kinds of tasks.

Turley says that, contrary to some reports, “I don’t think there’s going to be such a thing as an OpenAI agent.” What you will see is “increasingly agentic functionality inside of ChatGPT,” though. “Our focus is going to be to release this stuff as gradually as possible. The last thing I want is a big bang release where this stuff can suddenly go out and do things over hours of time with all your stuff.”

“The last thing I want is a big bang release”

By ChatGPT’s third anniversary next year, OpenAI will probably look a lot different than it does today. The company will likely raise billions more dollars in 2025, release its next big “Orion” model, face growing competition, and have to navigate the complexity of a new US president and his AI czar.

Advertisement

Turley hopes 2024’s version of ChatGPT will soon feel as quaint as AOL Instant Messenger. A year from now, we’ll probably laugh at how basic it was, he says. “Remember when all we could do was ask it questions?”

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Technology

Waymo raises $16 billion to take its robotaxi business ‘global’

Published

on

Waymo raises  billion to take its robotaxi business ‘global’

Waymo announced a $16 billion investment round aimed at bringing its robotaxi business to more US cities, as well as some overseas markets. The funding round was led by Dragoneer Investment Group, a “crossover” firm known for investing in late-stage tech companies before they go public.

Waymo’s co-CEOs said in a blog post they would use some of the money to buy more vehicles to grow its fleets size, a crucial step as it seeks to launch in at least 20 new cities in 2026. The company currently operates more than 2,500 robotaxis in six US cities. The new funding values Waymo at $126 billion.

Waymo’s latest funding round attracted several new investors, including Dragoneer, Sequoia Capital, and DST Global. Returning investors include Andreessen Horowitz, Abu Dhabi sovereign fund Mubadala, Fidelity Management and Research Company, Perry Creek Capital, Silver Lake, Tiger Global, Temasek, and T. Rowe Price. The company last raised a $5.6 billion in 2024, valuing the company at $45 billion.

Despite their promise to bring down costs by eliminating driver jobs, autonomous ridehail vehicles are enormously expensive. In addition to vehicle purchases, companies must install expensive sensors and computers into each vehicle. The robotaxis need to be monitored by remote operators during trips. And fleet managers handle EV charging, cleaning, and sensor calibration while the robotaxis are offline.

Still, Waymo is one of the few companies to run a paid service with fully driverless vehicles in the US. Amazon’s Zoox is still running free trips in a handful of cities, while Tesla has yet to transition away from using safety monitors in the vehicle.

Advertisement
Continue Reading

Technology

Super Bowl scams surge in February and target your data

Published

on

Super Bowl scams surge in February and target your data

NEWYou can now listen to Fox News articles!

The Super Bowl is not only the biggest sporting event of the year, but it has also become one of the busiest scam seasons.

Every February, millions of Americans receive texts, emails and calls tied to the game, such as “Your ticket couldn’t be delivered,” “Your streaming account needs verification” or “Your betting account was locked.” At first glance, these messages may seem like random spam, but in reality, they are carefully targeted.

Instead of blasting messages blindly, scammers rely on data brokers – companies that collect, package and sell personal information. These brokers build detailed profiles, and scammers either buy or steal those lists to decide exactly who to target and when.

Below, I’ll explain how this system works and, more importantly, how you can remove yourself from the data pipeline scammers depend on.

Advertisement

Sign up for my FREE CyberGuy Report
Get my best tech tips, urgent security alerts and exclusive deals delivered straight to your inbox. Plus, you’ll get instant access to my Ultimate Scam Survival Guide – free when you join my CYBERGUY.COM newsletter.

TAX SEASON SCAMS SURGE AS FILING CONFUSION GROWS

Scammers ramp up Super Bowl-themed fraud each February, using fake ticket, streaming and betting alerts to steal personal information from unsuspecting fans. (Elisa Schu/picture alliance via Getty Images)

Why Super Bowl season is a gold mine for scammers

Big events create urgency, emotion and distraction – perfect conditions for fraud. During Super Bowl week, scammers use the same themes real companies use:

  • Ticket confirmations
  • Streaming service alerts
  • Betting account warnings
  • Delivery delays for food or merchandise.

But here’s the key: they don’t blast these messages randomly. They target people who look like likely buyers. That targeting comes from your digital profile.

How scammers know you’re a ‘Super Bowl target’

You might never have posted about football, yet you still receive a fake ticket message. That’s because data brokers build profiles using your:

Advertisement
  • Address history
  • Household size
  • Income range
  • Age
  • Shopping behavior
  • Most-used apps
  • Frequently visited websites
  • Household members.

These details are stitched together from retailers, apps, public records and tracking tools – then sold to marketers and, eventually, leaked or resold to scammers. So when Super Bowl season arrives, scammers simply filter their lists: “People who look like they’d watch the game, place a bet, or order food.” And your phone number is right there.

The most common Super Bowl scam messages

During Super Bowl week, scammers flood inboxes and phones with messages that look like they came from legitimate companies. The goal is simple: create urgency, make you click and steal your information before you have time to think. Here are the scams I see spike every February:

1) Fake ticket alerts

“Your Super Bowl ticket transfer failed. Verify now.”

These messages pretend to come from Ticketmaster, StubHub or SeatGeek. They claim your ticket couldn’t be delivered, your transfer is pending or your account needs verification. The link takes you to a fake login page that looks identical to the real site. The moment you enter your email and password, scammers capture your credentials. Many victims then find their real ticket accounts emptied, their payment methods used or their email taken over.

How to spot it:

  • The sender’s address is misspelled
  • The link leads to a lookalike domain
  • The message creates panic and urgency.

What to do: Never click. Go directly to the ticket site through your browser or app.

2) Streaming account warnings

“Your Super Bowl stream is on hold. Update billing now.”

Advertisement

These messages impersonate major streaming platforms like YouTube TV, Hulu, ESPN, Peacock or cable providers. They claim your payment failed or your account is suspended just before kickoff. The link sends you to a fake billing page that steals your credit card details, login credentials or both. Some versions install malware that records keystrokes and login activity.

Why this works: Millions of people stream the Super Bowl. Scammers know most recipients won’t even question it.

What to do: Open the streaming app directly and check your account there. Ignore any links in the message.

5 MYTHS ABOUT IDENTITY THEFT THAT PUT YOUR DATA AT RISK

Super Bowl scams aren’t random; criminals use detailed data broker profiles to target likely viewers, bettors and shoppers. (Kim Kulish/Corbis via Getty Images)

Advertisement

3) Betting account freezes

“Your wager is pending. Confirm your identity.”

These target people who have been flagged by data brokers as likely sports bettors. Messages claim your account with DraftKings, FanDuel or BetMGM is locked due to “suspicious activity.”

The fake verification page asks for:

  • Your full name
  • Date of birth
  • Social Security number
  • Bank or card details.

This gives scammers everything they need to commit identity theft.

What to do: Never respond to betting account messages outside the official app.

4) Merch and food delivery scams

“Your Super Bowl order is delayed. Track here.”

Advertisement

Scammers mimic popular retailers and delivery apps, like Amazon, DoorDash, Uber Eats, FedEx and USPS. They claim your food, jersey or party supplies couldn’t be delivered.

Clicking the tracking link can:

  • Install malware
  • Redirect you to a fake login page
  • Steal your payment info.

Why it works: People are expecting packages and food orders that week, so the message feels real.

What to do: Use the retailer’s official app or website to check orders.

Why families are hit even harder

Your data isn’t isolated. Data brokers connect people living at the same address, spouses, children and roommates. So one exposed profile becomes an entire household target. During Super Bowl weekend, when everyone’s using phones, scanning QR codes and ordering food, one bad click can put the whole family at risk.

The real problem: you’re still on the lists

Most people try to protect themselves by:

Advertisement
  • Deleting emails
  • Blocking numbers
  • Installing antivirus software.

Those help, but they don’t stop your data from being sold again tomorrow. As long as your information exists in data-broker databases, scammers can keep finding you. That’s why I recommend removing your data at the source.

The ‘game-day cleanup’ that stops the targeting

If you want fewer scam messages, not just better filters, you need to remove your personal data from data brokers. That’s where a data removal service comes in. While no service can guarantee the complete removal of your data from the internet, a data removal service is really a smart choice. They aren’t cheap, and neither is your privacy. These services do all the work for you by actively monitoring and systematically erasing your personal information from hundreds of websites. It’s what gives me peace of mind and has proven to be the most effective way to erase your personal data from the internet. By limiting the information available, you reduce the risk of scammers cross-referencing data from breaches with information they might find on the dark web, making it harder for them to target you.

Check out my top picks for data removal services and get a free scan to find out if your personal information is already out on the web by visiting Cyberguy.com.

Get a free scan to find out if your personal information is already out on the web: Cyberguy.com.

What to do before Super Bowl weekend

Here’s how to protect yourself right now:

  • Don’t click Super Bowl messages. Even if they look real, go directly to the company’s website instead. Use strong antivirus software to help block malicious links, fake websites and malware before they can steal your information.Get my picks for the best 2026 antivirus protection winners for your Windows, Mac, Android and iOS devices at Cyberguy.com.
  • Avoid QR codes from emails or texts. Many link to fake login pages.
  • Use credit cards, not debit cards. They offer stronger fraud protection.
  • Remove your data from broker sites. This is the single most effective way to reduce scam targeting.

FBI WARNS QR CODE PHISHING USED IN NORTH KOREAN CYBER SPYING

Fake Super Bowl tickets and streaming messages often create urgency, pushing victims to click malicious links or fake login pages. (Kyle Ericksen/Penske Media via Getty Images)

Advertisement

Kurt’s key takeaways

Super Bowl scams are not random. Instead, they are precision-targeted using personal data sold behind the scenes. While you cannot stop scammers from trying, you can make it harder for them to find you. By removing your data now, you reduce scam messages, limit fake alerts and lower your risk, not only this Super Bowl, but throughout the entire year. That kind of protection is a win worth celebrating.

Have you received scam texts or emails tied to the Super Bowl? What did they look like, and did you almost fall for one? Let us know by writing to us at Cyberguy.com.

CLICK HERE TO DOWNLOAD THE FOX NEWS APP

Sign up for my FREE CyberGuy Report
Get my best tech tips, urgent security alerts and exclusive deals delivered straight to your inbox. Plus, you’ll get instant access to my Ultimate Scam Survival Guide – free when you join my CYBERGUY.COM newsletter. 

Copyright 2026 CyberGuy.com. All rights reserved.

Advertisement

Continue Reading

Technology

Shedding light on Iran’s longest internet blackout

Published

on

Shedding light on Iran’s longest internet blackout

After protests broke out in early January, the Iranian regime shut down the internet, starting the longest blackout in Iranian history. Despite this attempt to stop the protests from spreading, they did not stop. Still, the internet shutdown slowed down the spread of information both inside and outside Iran.

Behind the heavily policed borders and the jammed signals, an unprecedented wave of state violence continues to add to a death toll somewhere between 3,000 and 30,000. Even at the lowest count, which has been acknowledged by the Iranian state and is likely a wild underestimate, these last few weeks have been one of the bloodiest uprisings in modern history.

The situation in Iran can be hard to grasp. The history is complicated; the state of the technology and internet infrastructure there is constantly in flux. To get a sense of what is happening right now, I turned to an expert. Mahsa Alimardani, the associate director of the Technology Threats & Opportunities program at WITNESS, has been a researcher and advocate in the digital rights space — particularly around Iran — since 2012. I spoke with her about what is happening in Iran, and how technology both props up and threatens repressive regimes.

The Verge: What is internet access in Iran like right now?

Mahsa Alimardani: Since the weekend [of January 24], there has been some resumption of connectivity. And I’m a little bit worried that this might convince people that things are back to normal. Last I saw, there was like 30 to 40 percent connectivity on some of the Cloudflare network data in Iran and there’s very inconsistent connectivity. Some circumvention tools have started to work.

Advertisement

Randomly, someone in Iran FaceTime called me yesterday. They were like, “My VPN stopped working, so I just tried to call with FaceTime, and for some reason, it didn’t even need a VPN.” But it was a momentary glitch. Various things are happening across the network, and it’s not really clear why there’s this opening, or what it means for long-term connectivity.

Since January 8th, when there was a surge in the uprising in the protest movement in Iran, there was an internet shutdown — the longest internet shutdown in Iran, they broke the record in length.

They also broke the record in number of protesters that have been massacred. It’s horrifying to think that technology helps enable such crimes.

Why does the Iranian government fear internet access?

In 1988, there was a fatwa where the government massacred a lot of political prisoners in a short span of time. I bring this up because it happened when there was no internet, and the media was heavily controlled and centralized by the state. If you did not flee Iran, and if you were not part of the generation of prisoners and political activists that survived, it was very hard to pass on the memory of that event. Peers of mine in Iran didn’t grow up with the same information. It’s so interesting having these conversations with people and realizing they are learning history only when they leave the country.

Advertisement

What’s been a real game changer is the way you can document and witness these kinds of crimes in the age of the internet. I think it’s obviously a big threat to the regime. It’s a massive threat to them to be able to hold them accountable, and be able to document and witness what they’re doing.

Anytime anyone sees a severe crackdown like an internet shutdown, you know that it’s going to be followed by violence. In 2019 there was a week-long internet shutdown, under the blanket of which they massacred 1,500 people. The reason why is because they don’t want people to use the internet for mobilization and communication, and they don’t want there to be a way to document what’s happening.

Anytime anyone sees a severe crackdown like an internet shutdown, you know that it’s going to be followed by violence

So the denial of the scale of their crimes is part of what they do in Iran, because it’s very hard to assess the percentage of legitimacy that the regime has, because obviously you can’t do free polling. You don’t have free media. Even when you have foreign journalists that go there, they’re followed by minders and the reporting is super-limited. The UN hasn’t been able to really have anyone do proper site visits for human rights documentation, since the start of this regime in 1979.

There isn’t any real access to professional on-the-ground documentation and fact-finding. So it all really depends on the internet, on people, on citizen media. People sending things, putting them online, and then having professional fact-checkers and verification.

Advertisement

What was internet access in Iran like most recently? What platforms and service providers did people use before the blackout started?

Iranians are extremely tech savvy because there’s been a cat-and-mouse game across the internet for most of its existence. Since 2017, 2018, on average, there’s been protests every two years. Each time they have a different level of censorship, new kinds of rules and regulations.

In 2017, [messaging app] Telegram was massive. Some people were even saying Telegram was the internet for Iranians, they were doing everything across Telegram. It worked really well, especially with network bandwidth being really low. So Telegram was a place for news, chatting, socializing everything, even like online markets. But then they blocked it in 2018 when protests started, because protest mobilization on there was a threat to the regime.

There was a move toward Instagram and WhatsApp becoming the most popular applications.

They had yet to be blocked back then. Instagram was more for fun, but it became much more politicized after Telegram was blocked. Then, during the Woman Life Freedom movement in 2022, Instagram and WhatsApp got blocked.

Advertisement

The regime has spent a lot of effort in trying to disable VPNs

Most people are just on VPNs. The regime has spent a lot of effort in trying to disable VPNs. There’s a lot of different VPN projects both for-profit and nonprofit that work within that cat-and-mouse game where protocols are being disabled and new ones are created.

An average Iranian often has many different VPNs. When one can’t work, they’ll turn on another one.

We’ve talked about how technology threatens the regime and how average Iranians use it. Let’s switch over to the other side of this issue: how does technology enable repression?

So there’s various different things the regime does, different levels of enacting information controls. There’s the censorship level of shutting it down.

Advertisement

Then there’s physical coercion. Like, I know people who have not reported their children who have been killed recently because they were so frightened by the process by which they had to get their loved one’s body.

They also flood the information space with a lot of misinformation. They create a lot of doubt.

They’ve been doing this information manipulation even before the internet. Iran is a very complicated information space. There are a lot of actors beyond the regime who also want to manipulate it. Even authentic dissidents and activism will get lumped in with Mossad or CIA operations.

Iran’s foreign relations muddy its information space

In 1953, the American CIA and British MI6 overthrew the democratic government of Iran, consolidating power under a monarchy that was more favorable to the US and the UK. Many believe that the political instability caused by the CIA and MI6 eventually led to the Islamic Revolution of 1979, which established the current authoritarian regime.

Advertisement

From 2014 to 2024, Iran and Russia joined a strategic partnership with the Syrian dictatorship as part of the Syrian civil war. The United States formed its own coalition; both coalitions purported to fight ISIS. The civil war spawned massive amounts of internet disinformation, and in 2018, Facebook and Twitter deleted hundreds of accounts originating in Russia and Iran that formed a global influence network pushing disinformation. The Syrian regime was overthrown at the end of 2024. The next year, following decades of hostilities, Israel and Iran engaged in a 12-day war.

These are some, but not all of the factors that contribute to the complicated information space in Iran that Alimardani is referring to.

The regime’s campaign existed pre-internet, but with technology, it went into overdrive. They’ve been quite clever in some of the ways they’ve covered the protests. They’ve been able to even mobilize, like, people who are sympathetic to the Palestinian cause, against, you know, the Iranian cause for liberation.

There have been a lot of documented efforts of them trying to manipulate protest documentation, undermine it, you know, use the concept of the Liar’s Dividend, which is very easy to use in the increasingly AI world we’re in.

Hold on, can you go through those examples you just mentioned? About mobilizing people who are sympathetic to the Palestinian cause?

Advertisement

Yeah, so, Iran is quite complicated in that it’s an Islamic fascist state. They use Islam in a lot of ways to repress the people. And there is a lot of very valid rhetoric about Islamophobia in the West, from the very specific context and history of the United States, such as what happened during the War on Terror.

But in Iran, it’s quite different. And this can really be manipulated and conflated, right? Mosques in Iran are often also the headquarters for the Basij [the Iranian paramilitary corps], and people might not know this. So there will be videos like, “Look at these protesters who are setting fire to this mosque. Look at these Islamophobic rioters.”

You might see that, without the context that the mosques also are places where the security forces that kill people are stationed, and lose why something like that would be attacked by Iranians seeking liberation.

You mentioned the regime’s use of AI — do you want to talk a little bit more about that?

Yeah, so, we didn’t need AI for authoritarian regimes to deny evidence of their crimes. Even before AI, Bashir al-Assad [the former dictator of Syria] was saying that reliable documentation of his crimes in Syria were not valid.

Advertisement

Whether we like it or not, AI is being integrated into a lot of things. AI editing is slowly becoming ubiquitous. Like, in fact, we might come to a point where editing photos or anything might become unavoidable without the use of generative AI.

So you no longer have that binary of like, if it’s AI, it’s fake. If there’s no AI, it’s real.

So there’s this very symbolic image that everyone has said reminded them of the Tiananmen Square Tank Man from 1989. But here, a protester is standing in front of armed security forces on motorcycles with weapons. [Ed. note: The New York Post ran with the headline “Powerful image of lone Iranian protester in front of security forces draws parallels to Tiananmen Square ‘Tank Man.’”]

This was a very low resolution video taken from a high rise [building]. Someone had screenshotted a frame from the video and it was quite blurry.

They used some AI editing software to enhance it, and you could see some AI artifacts. Nevertheless, this is an authentic, verified image of a brave protester. Lots of credible sources have verified it. But immediately, it was pointed out to have these AI artifacts, and a lot of the regime accounts started this narrative of “This is all AI slop from Zionists.”

Advertisement

And of course, because, you know, Israel has a special interest in Iran, they have a Farsi-language state account. Israel’s Farsi state account shared the image, which further fueled the claim that this authentic image from Iran was AI slop being pushed by the enemy, Israel.

As you’ve already mentioned, Iran has a complicated information environment. What would you say are the various actors in this space? What kinds of things are they doing?

Obviously there are foreign policy interests by Israel and the US in Iran, just because of the history and very antagonistic relationship they’ve had from the very beginning of the revolution.

The Iran-Israel war in June 2025 was a super interesting moment because the war started a few weeks after Google launched Veo 3, which has made access to very realistic generative AI content very easy. So right off the bat, you could see, from both sides, a lot of AI content coming from the war. This wasn’t the first war where that’s happened — like the Ukraine war has had so many different examples — but since Russia’s invasion of Ukraine [beginning in 2022], the technology has advanced far more, so it became a very big part of the narrative of the situation in Iran.

The most famous example from the Iran-Israel war was a piece of manipulated content that Citizen Lab later was able to attribute to the Israeli state. It was this AI-generated video of Israel bombing the gates of Evin Prison, perpetuating this narrative that they have very precise military operations and that they were freeing these political prisoners.

Advertisement

Evin is a very famous prison for a lot of activists and dissidents and intellectuals in Iran. Human Rights Watch and Amnesty International called the bombing of Evin Prison a war crime. And indeed, political prisoners were casualties of the bombing.

But that deepfaked video went viral. Mainstream media even reposted it immediately before a lot of various different researchers, including our deepfakes rapid response force and others, were able to attest that indeed this was a manipulated video.

So you have this information space that is quite complicated. But in this scenario, I think it would really be remiss to put that much emphasis on the role that these other actors have. There are things from these outside actors that fog up the information space, but ultimately what’s really happening is that there’s a really unprecedented massacre happening. And the perpetrator is the Islamic Republic of Iran.

I’ve seen some reporting about how Iranians bought Starlink terminals prior to the blackout. Can you say anything about that?

Yeah, I want to start by referencing a really great article by the Sudanese activist Yassmin Abdel-Magied, called “Sudanese People Don’t Have the Luxury of Hating Elon Musk.” Whatever my personal ideas are about Elon Musk, you have to give credit where credit is due. This technology is a game changer. It’s been a game changer in Sudan. And it has been in Iran.

Advertisement

We’ve had a few days of a little bit of connectivity of people coming online just through the ordinary network, but when the shutdown was full and complete, Starlink was really the only window we had into Iran.

When the shutdown was full and complete, Starlink was really the only window we had into Iran

And if you talk to documentation organizations, they’ll tell you, they were getting evidence and doing the verification through what was coming in from the Starlink connections. I know of people who had a Starlink and had like a whole neighborhood of people come in to check in and use the Wi-Fi.

The most credible stats before the situation was that there’s about 50,000 Starlinks. There’s likely more than 56,000 now. It became very popular during the Iran-Israel War, because of course, then the Islamic Republic enacted another shutdown. A lot of people invested in getting Starlink then.

You can get anything you want in Iran through smugglers — I think Starlink was like $1,000 at the time because demand was so high. Receivers are ordinarily a few hundred US dollars. The last price I heard was they were being sold for $2,000 in Iran. It’s a lot of money, but given the demand and the massive risk the smugglers have to undertake, I think it’s fair, but also, it means you can’t really scale this, and the people that have it are very privileged or have access to very privileged people.

Advertisement

What we’re seeing is a very small window. When having discussions with various folks that have been doing firsthand documentation, they’ve expressed, “We’re not getting enough from Kurdistan. We’re not getting enough documentation from Sistan and Baluchestan.” Historically, these areas are often at the forefront of protests, because the regime often has the bloodiest forms of repression in these provinces with marginalized ethnicities. Areas like Sistan and Baluchestan have a lot of economic poverty, so they’d have less access to something privileged like Starlink.

Satellite internet is really this way of reimagining connectivity

For all these years, myself, many people, have been working on this concept of internet censorship and internet shutdowns. And there really hasn’t been a way to reimagine this system. There’s this concept of digital sovereignty in place in terms of internet access and internet infrastructure that fits within national borders. In even the most democratic of countries, this is still national infrastructure that the government can have access or forms of control over.

This concept has to be broken. Satellite internet is really this way of reimagining connectivity, not just for Iran, but anywhere where lack of connectivity results in a crisis, whether humanitarian one, or a massacre of this proportion.

It’s really important to reconceive access to satellite internet in a way that could scale beyond those who are privileged and beyond those willing to take the risk. And one of the ideas that I’ve had and have been working on with other colleagues at Access Now has been to push for direct-to-cell access, which is a form of satellite internet connectivity that depends on technology that exists in phones created from 2020 onwards. We launched this campaign called Direct 2 Cell, hoping to push forward this concept.

Advertisement

On a personal note, how are you doing? Have you heard from your friends, family, other people you know in Iran recently?

I’ve been able to be in touch with some of my family and others here and there.

I also had that random FaceTime audio call from another person I know. I was very worried about them because they’ve been at the protests. I had heard through various people that they were okay, but I finally heard from them firsthand, and it was such a bizarre experience, speaking to them.

I had never heard them sound the way that they sounded: recounting their experience of leaving the protest before the military tanks came to open fire on the crowds, how they got tear gassed, and for the next few days, seeing water hoses washing blood off the streets. It sounded like they were making a lot of dark jokes — I had never heard them sound this way. I don’t know how you can walk the streets of your neighborhood, seeing people wash off blood, and just…. like, something not fundamentally change in your mind.

I just, I don’t, I can’t imagine how to process it if I was there. As someone in the diaspora, it’s hard to process being privileged and being away.

Advertisement
Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.

Continue Reading

Trending