# Why Deploy Abuse Detection If It's Not 100% Accurate?

Tisane offers high accuracy, but like all natural language processing systems, it is not 100% perfect. Nor are human moderators.

## Why Use It Anyway?

1. Deterrence Against Offenders
  - Just as law enforcement cannot be everywhere, moderation tools do not need to be perfect to be effective.
  - The mere presence of an abuse detection system discourages bad actors from engaging in harmful behavior.
2. Speed and Efficiency
  - NLP systems can analyze large volumes of content instantly, something human moderators cannot scale to match.
  - Even if some cases require human review, the system significantly reduces the workload.
3. Transparency and Consistency
  - Automated moderation ensures rules are applied fairly, reducing bias and human inconsistency.
  - Unlike human moderators, NLP systems do not get tired, emotional, or overwhelmed.
4. Real-World Comparisons
  - No police force prevents all crime, yet they reduce incidents and deter criminals.
  - No spam filter catches 100% of spam, yet they dramatically improve email usability.
  - No lock is 100% tamper-proof, yet it's much safer to keep the door locked than leave it open at all times.
5. Continuous Improvement
  - Tisane Labs is committed to improving accuracy over time.
  - Real-world usage helps fine-tune detection models and reduce false positives and negatives.


For reference, even serious crimes often remain unsolved ([source](https://www.reddit.com/r/UnresolvedMysteries/comments/5toay8/what_percentage_of_murders_are_unsolved/)), yet law enforcement remains essential. Similarly, abuse detection tools are valuable even if they aren’t perfect.