Exactly. Tor was originally created so that people in repressive countries could access otherwise blocked content in a way it couldn’t be easily traced back to them.
It wasn’t designed to protect the illegal activities of people in first world countries that have teams of computer forensics experts at dozens of law enforcement agencies that have demonstrated experience in tracking down users of services like Tor, bitcoin, etc.
Oh there are definitely ways to circumvent many bot protections if you really want to work at it. Like a lot of web protection tools/systems, it’s largely about frustrating the attacker to the point that they give up and move on.
Having said that, I know Akamai can detect at least some instances where browsers are controlled as you suggested. My employer (which is an Akamai customer and why I know a bit about all this) uses tools from a company called Saucelabs for some automated testing. My understanding is that our QA teams can create tests that launch Chrome (or other browsers) and script their behavior to log into our website, navigate around, test different functionality, etc. I know that Akamai can recognize this traffic as potentially malicious because we have to configure the Akamai WAF to explicitly allow this traffic to our sites. I believe Akamai classifies this traffic as a “headless” Chrome impersonator bot.
When any browser, app, etc. makes an HTTP request, the request consists of a series of lines (headers) that define the details of the request, and what is expected in the response. For example:
GET /home.html HTTP/1.1
Host: developer.mozilla.org
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9; rv:50.0) Gecko/20100101 Firefox/50.0
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8
Accept-Language: en-US,en;q=0.5
Accept-Encoding: gzip, deflate, br
Referer: https://developer.mozilla.org/testpage.html
Connection: keep-alive
Upgrade-Insecure-Requests: 1
Cache-Control: max-age=0
The thing is, many of these headers are optional, and there’s no requirement regarding their order. As a result, virtually every web browser, every programming framework, etc. sends different headers and/or orders them differently. So by looking at what headers are included in a request, the order of the headers, and in some cases the values of some headers, it’s possible to tell if a person is using Firefox or Chrome, even if you use a plug-in to spoof your User-Agent to look like you’re using Safari.
Then there’s what is known as TLS fingerprinting, which can also be used to help identify a browser/app/programming language. Since so many sites use/require HTTPS these days it provides another way to collect details of an end user. Before the HTTP request is sent, the client & server have to negotiate the encryption to use. Similar to the HTTP headers, there are a number of optional encryption protocols & ciphers that can be used. Once again, different browsers, etc. will offer different ciphers & in different orders. The TLS fingerprint for Googlebot is likely very different than the one for Firefox, or for the Java HTTP library or the Python requests package, etc.
On top of all this Akamai uses other knowledge & tricks to determine bots vs. humans, not all of which is public knowledge. One thing they know, for example, is the set of IP addresses that Google’s bots operate out of. (Google likely publishes it somewhere) So if they see a User-Agent identifying itself as Googlebot they know it’s fake if it didn’t come from one of Google’s IP’s. Akamai also occasionally injects JavaScript, cookies, etc. into a request to see how the client responds. Lots of bots don’t process JavaScript, or only support a subset of it. Some bots also ignore cookies, and others even modify cookies to try to trick servers.
It’s through a combination of all the above plus other sorts of analysis that Akamai doesn’t publicize that they can identify bot vs human traffic pretty reliably.
Exactly. The only truly effectively way I’ve ever found to block bots is to use a service like Akamai. They have an add-on called Bot Manager that identifies requests as bots in real time. They have a library of over 1000 known bots and can also identify unknown bots built on different frameworks, bots that impersonate well known bots like Googlebot, etc. This service is expensive, but effective…
Not easily. The scammer likely has your current address & contact info, but knows nothing about your history.
To confirm your identity when you contact these reporting agencies they will use details from your credit history by asking detailed questions the scammer likely won’t know. For example it might be questions like these:
- What kind of car did you purchase in 2005?
- Honda
- Ford
- Saab
- Jeep
- None of the above
- Which one of these companies did you work for previously?
- IBM
- Pizza Hut
- Macy’s
- Jiffy Lube
- None of the above
They’ll throw 3 or 4 questions like these at you that you’ll have to answer correctly. They might involve places you used to live, banks you have had accounts with, etc. The chances of a scammer with your SSN knowing all these details about you is pretty tiny.
The credit monitoring companies have your up-to-date contact information (and verified) when you put the freeze in place. Now, should a third party try to open an account, etc. in your name it should be blocked from happening and the credit monitoring company should contact you.
If a scammer tries to unfreeze or otherwise modify your account with them they should also contact you.
If/when they contact you or you request your account be unfrozen then they’ll use old credit history to confirm your identity. These are a series of three or four random questions that a scammer is unlikely to know. For example they might ask you what kind of car you purchased in 2005, then give you 4 options, like Ford, Honda, Jaguar, or BMW, and then also a “nine of the above” option. Then they might ask you which of the following street addresses you used to live at, and list 4 seemingly random addresses, one of which you might have lived at.
robots.txt is 100% honor based. Well known bots like Googlebot, Bingbot, etc. definitely honor them. But there are also plenty of bots that completely ignore them.
I would hope the bots used to collect LLM training data honors them, but there’s no way to know for certain. And all it really takes is one bot ignoring it for the content of your website to end up in a random set of training data…
Hate to break it to you but all the major CDN providers do the exact same things. My employer runs multiple websites mainly for US and European users. We use Akamai for both CDN and WAF services. For any CDN and/or WAF to operate properly it needs access to unencrypted content. Part of Akamais WAF tools includes what they call Bot Manager, which can identify traffic coming from over 1000 known bots and can also classify unknown ones. Part of how it works is by browser fingerprinting as well as TLS session fingerprinting and other proprietary fingerprinting.
So any time you visit a large website you’re likely being fingerprinted and otherwise analyzed by the CDN and security tools used by those sites.
It’s a laptop owned by your school, so they can install spyware if they want to. More importantly the school likely has policies against removing or otherwise tampering with it. You would be wise to find out what they will do if you violate this policy. It could be anything from a slap on the wrist to expulsion.
Any decent IT department will eventually figure out if you disable it. They’ll know fairly quickly if it stops “phoning in” if the spyware is any good.
Well OPSEC is the stated cause. Who knows how the person was initially identified and tracked. For all we know he was quickly identified through some sort of Tor backdoor that the feds have figured out, but they used that to watch for an unrelated OPSEC mistake they could take advantage of. That way the Tor backdoor remains protected.