The problem

I won't point fingers at people, I won't name any names, but many of these creators and tutorials involve payment methods and sensitive data which you will put in the AI chat, and that alone raises security alarms everywhere.

All it takes is for 1 hacker / intruder to gain access to your chat, and you've lost all your money, your personal data is all over the world, and it spirals out of control from there. Not only that, but AI in general is untrustworthy. In almost every single AI model in the world, it will have a message somewhere stating something along the lines of "xxxx is an AI model, and may get things wrong". Now imagine if it makes a mistake that leaks all of your data without you even having to lift a finger and you wouldn't even know. 1 data breach and again, your personal / private data is now publicly available.

I haven't even listed much of anything, these are just 2 very possible outcomes that could happen very easily, and just remember:

Your data is private for a reason!!

Here, I've used OpenClaw as an example, but this happens with many different tools, models, etc. OpenClaw is just currently very popular, and there are tons of real life horror stories / cases about this that are just 1 google search away.

What does this mean?

Anthropic even has their own safety guide for this extension, which tells us 2 things:

• Anthropic realises the danger which is opposing them, and are taking action to try to prevent data breaches

• Your data is completely out of their hands, and there's no humans except from you, the user to control what the AI does.

This can be perceived either in a positive way, as they genuinely care about your data, or as the harsh reality of AI, the fact that it almost has it's own brain, and therefore it's actions are completely unpredictable.

On the contrary, Claude for Chrome can be used alongside Claude Code, this introduces the idea of Claude being a complete developer on it's own. If you removed its permission limits, and let it go free, I'm 99% sure that it could create a whole new website using Umbraco for example, make it fully interactive, and with all of the right tools and resources, it could most likely push said site live to the world, all from scratch. Of course this doesn't mean you can use this instead of real developers, as you still need that human to fact check everything, and AI, as mentioned previously, is untrustworthy on it's own. 

What does this mean for you?

This could mean that while you are working on a client site for example, you could set claude off to create you a site from scratch exactly how you want, and by the end of the day, you could have finished all of your client work, and also have a fully immersive site for yourself.

What does this mean for illegitimate people?

While this is great for the average Joe Bloggs, this imposes some threats. The one that stands out above all for me would be mass production of illegitimate sites, or anything along those lines. This is because if I were to buy 100 domains, and opened 10 Claude instances, and I wired my own malicious code into the claude chat, I could make it write me 100 sites that look extremely professional, and exactly what you (the user) need, but in reality, it is full of Malware. 

This is already a huge issue, and it is called a "Trojan attack", however currently, the malicious user would need to create the sites manually, and fill the content, code, etc on their own, which would take a while to get a lot of sites and content out there. Whereas now, if my logic is correct they can automate this fully to create tons of sites, all malicious, all dangerous.

Overview

This post isn't meant to make you scared of AI, or the internet, but rather to raise awareness. Too many people are trying to use AI without taking the necessary precautions and having the knowledge that is needed, and that a a result is becoming weaponised.

Whether you like it or not, there will always be someone somewhere willing to take your data for one reason or another, so the next time you use AI, or anything on the internet really, just think to yourself:

"Am I doing everything in my power to protect myself?"

If the answer is no, then take a step back, and try it again.

This post isn't to deteriorate you from using tools like the ones mentioned either. As long as you have done everything in your power to protect yourself, you will be safe most of the time, and no matter what you are told or think:

NEVER share your personal / sensitive data with AI