Regarding Facebook's Risk Control Mechanism

[baishiyi92020]

I obtained plaintext requests for Facebook and the app, and modified numerous parameters, yet risk control measures still triggered. I'm unsure where else information is being correlated, preventing me from performing bulk registrations.
Is it because they've integrated an AI risk control system that can perform data analysis on our registration process?

 

[DQ Agency]

Most likely yes. Facebook now checks many signals like IP reputation, device fingerprint, behavior patterns, and registration speed. Even if the request looks clean, those other signals can still trigger risk control.

 

[baishiyi92020]

Most likely yes. Facebook now checks many signals like IP reputation, device fingerprint, behavior patterns, and registration speed. Even if the request looks clean, those other signals can still trigger risk control.

Yep, I know it checks multiple items, so I've already modified the script. The IP should also be a decent proxy, as it currently allows registering several accounts daily before subsequent registrations trigger risk controls.

 

[Jenny 15]

I obtained plaintext requests for Facebook and the app, and modified numerous parameters, yet risk control measures still triggered. I'm unsure where else information is being correlated, preventing me from performing bulk registrations.
Is it because they've integrated an AI risk control system that can perform data analysis on our registration process?

View attachment 511201

Absolutely, Facebook has an AI-powered risk control system that can analyze data during your account registration process. For example, you might sometimes encounter accounts being simultaneously deactivated for the reason "risk payment".

 

[baishiyi92020]

Absolutely, Facebook has an AI-powered risk control system that can analyze data during your account registration process. For example, you might sometimes encounter accounts being simultaneously deactivated for the reason "risk payment".

Holy shit, bro, is there any way to fix this? I can modify all the requests.

 

[duonght12]

I obtained plaintext requests for Facebook and the app, and modified numerous parameters, yet risk control measures still triggered. I'm unsure where else information is being correlated, preventing me from performing bulk registrations.
Is it because they've integrated an AI risk control system that can perform data analysis on our registration process?

View attachment 511201

Yes, Facebook now uses advanced automated systems (including AI/ML) for risk detection. They don’t rely only on request parameters.
The system can also analyze things like device fingerprints, IP reputation, browser behavior, cookies, account patterns, and other signals during the registration process.
That’s why simply modifying request parameters usually isn’t enough anymore. The platform evaluates the overall behavior and environment to detect suspicious activity.

 

[baishiyi92020]

Yes, Facebook now uses advanced automated systems (including AI/ML) for risk detection. They don’t rely only on request parameters.
The system can also analyze things like device fingerprints, IP reputation, browser behavior, cookies, account patterns, and other signals during the registration process.
That’s why simply modifying request parameters usually isn’t enough anymore. The platform evaluates the overall behavior and environment to detect suspicious activity.

However, since I can modify all parameters, it means the data he receives is exactly this. So what I need to do now is identify the risk control parameters that add points, right? The IP issue is mostly resolved, and I can modify device parameters. But currently, I can't precisely determine which parameters are high-risk because the same parameter might work one time and fail the next.

 

[aprildfwida]

That’s very possible. Facebook now evaluates many different signals such as IP reputation, device information, behavior patterns, and how quickly an account is created or used. Even if everything appears normal on the surface, those background signals can still trigger additional security checks or risk reviews.

 

[baishiyi92020]

Yeah it’s very likely not just request parameters anymore, because modern platforms correlate signals far beyond API data, including device fingerprints, behavioral timing, network reputation, sensor data, and long-term pattern analysis that’s hard to spoof at the request level alone. What people often interpret as “AI risk control” is really multi-layered trust scoring where registration flow consistency, environment realism, and historical signals matter more than modifying payloads themselves.

So why do merchants still have so many accounts in stock?

 

[baishiyi92020]

Okey,My theory is sound; adjusting the parameters can mitigate the risk control issues.