Serendipity35

 

YouTube is giving all creators 18 and over access to a tool that can detect whether their likeness has been copied and used in AI videos uploaded to the website. Known as deepfakes, these are AI-generated videos, images, or audio that make it look like someone said or did something they never actually said or did.

Team YouTube made an announcement, explaining that their "goal is to provide [users] with more peace of mind by giving [them] easy access to request the removal of unauthorized content." While their likeness detection tool is technically available only to creators, spokesperson Jack Malon said that anyone can use it. 

The tool allows you to:
Identify where your likeness shows up.
Understand if others are using your face in videos made with AI. 
Safeguard your identity by requesting removal of unauthorized content that uses your likeness directly in YouTube Studio.
Keep your viewers from being misled about you.

An AI model is trained on lots of real photos, videos, or audio of a person — the more data, the better the fake. Then, using techniques like generative adversarial networks (GANs) or diffusion models, the AI swaps faces, clones voices, or creates entirely new footage that mimics the person’s appearance, expressions, and voice. Modern tools can match lighting, lip movements, blinking, and even subtle mannerisms to make the result really convincing.

In a face swap, it replaces one person’s face with another face. A lip sync can change what someone appears to be saying
Voice cloning can copy someone’s voice from audio samples. What is known as full-body puppetry can control gestures and expressions and make a still photo “talk” and move.

Most articles about this technology are negative and point out risks, such as misinformation, non-consensual fake content, scams, identity fraud, and reputation damage.

If you do see positive uses, they will often not be called "deepfakes." The positive uses are still controversial but include the use of the technology in filmmaking, dubbing movies into other languages without reshoots, some accessibility tools, satire, education, and bringing historical figures “back to life.”

Some platforms like Meta's Facebook, Instagram, and Threads now label or remove certain deepfakes. 

A well-known example of a deepfake is the 2018 fabricated video of former U.S. President Barack Obama created by comedian and director Jordan Peele. Using artificial intelligence, Peele superimposed his own voice and facial movements onto a digitally replicated face of Obama, successfully demonstrating how AI can be used to put words into anyone's mouth.

YouTube announcement at https://support.google.com/youtube/thread/434105667

 

In April, the makers of Claude shared that the company's new AI model (Mythos 5) was too dangerous to release to the public. Weeks later, thanks to some tweaks, the new model (called Fable 5) was released to the public. Now, it has been announced that "Anthropic has suspended its powerful new AI model after US authorities raised security concerns just days following its public release."

I find some irony in these AI names. Mythos and Fables indeed.

MORE    Anthropic's Claude Fable 5 and Mythos 5 AI suspended over security fears  

This month, colleges and universities across the country postponed final exams and due dates for assignments after Canvas, a learning management system used by 41 percent of North American higher ed institutions, temporarily went offline due to a hack. The University of Illinois at Urbana-Champaign postponed “all final exams and assignments, including papers, projects, etc., scheduled for Friday, Saturday, or Sunday,” provost John Coleman wrote to students and employees, and that, for “consistency and clarity,” the postponement affects all classes—even those that don't use Canvas.

Cybercrime group ShinyHunters identified itself as the hackers.

ShinyHunters first emerged in 2020 and claims to have successfully attacked 91 victims so far. The group is primarily after money, but has also been willing to cause reputational damage to their victims. In 2021, ShinyHunters announced they were selling data stolen from 73 million AT&T customers. ShinyHunters received global attention in 2025 after Google urged 2.5 billion users to tighten their security following a data breach via Salesforce, a customer management platform.

Unlike data breaches where hackers directly break into databases holding valuable information, ShinyHunters – and several other groups – have recently targeted major companies through voice-based social engineering, which is also known as “vishing,” for voice phishing. Social engineering is when a person is tricked or manipulated into providing information or performing actions that they wouldn’t normally do.

The phrase that if an app is free, you are the product means that when an app doesn’t charge you money, it usually makes money from you instead. They do that mainly by collecting your data or selling your attention to advertisers.

If that is true, then how is AI changing what that means? It is a question that deserves several posts here to really answer.

Your behavior, preferences, and time become what is being monetized. Your data becomes the product. Free apps often gather your demographics, browsing or in-app behavior, location, interests, and habits. This information is then used to target ads or sold to third parties.

The addictive nature of app design keeps you scrolling, tapping, or watching so they can show you ads. You pay with time, not dollars. “Free” is a business model, not a gift.

I will give these companies a nod that running an app costs money (servers, engineers, storage). If you are not paying, the company must earn revenue another way. Ad-free options are becoming more common as a premium. You have probably noticed that on apps and also on video streaming services. You thought that paying for Amazon Prime meant no ads on the videos. Wrong. Free is often an illusion.

In the world of AI, the difference between free and paid tiers is more than a matter of convenience. It is also about identity and privacy.

Privacy becomes the hidden cost. Data is currency. Companies track you across apps and devices, build detailed behavioral profiles, and use algorithms to influence what you see. This raises concerns about autonomy and consent.

Is there no stopping them? As long as you agree to their terms, they have a lot of power. BUT you can read those terms and privacy settings more carefully. (They rely on the fact that many users don't read the terms or adjust their settings at all.) Educate yourself and understand how digital ecosystems make money. You can choose paid or privacy-focused alternatives. And you can remove the app entirely from your life.

I see comparisons of using AI to using social media platforms. I don't think AI data is the same as social media data. Social media platforms monetize your attention. The longer you scroll, the more ads they can show. AI chatbots operate on a different axis. Your prompts aren’t just content; they’re training signals. They reveal how people think, what they struggle with, what they’re curious about, and how they phrase questions. Maybe it is anonymized (a good thing) but it is still valuable and often sensitive data.

Alarmist articles will remind you that many free AI chatbots use your prompts, your corrections, and your uploaded files. They have that photo of your family that you let them enhance. What will they do with what you give them? I can't answer that as of now, and certainly not for the future. I know that your conversation history is used to train or fine-tune future versions of the model. Hey, you are part of the product pipeline - but don't expect to be paid for your contributions.

I also concede that the business model matters and that different AI companies monetize differently. For example, Microsoft provides its own privacy commitments and policies, and those govern how your data is handled. For details, they always direct users to their Privacy Statement.

Here are 4 business models currently out there:
Ad-supported = Your attention is monetized.
Freemium = Free tier gathers usage; paid tier subsidizes development.
Enterprise licensing = Your data may be isolated; the company earns from businesses.
Open source =  The model is free; the company may sell hosting or support.

So "if an app is free, you are the product" still applies, but not always in the same way. When an AI tool is free, you’re not just the product — you’re also the collaborator. You’re an unpaid teacher, tester, and a source of fuel for improvement.