diff --git a/adlib/tests/README.md b/adlib/tests/README.md
new file mode 100644
index 0000000000..d83beeacce
--- /dev/null
+++ b/adlib/tests/README.md
@@ -0,0 +1,16 @@
+# Data Poisoning Tests
+
+This README is designed to give a quick overview about data poisoning (DP)
+adversaries, learners, and tests in the `adlib` library. The available
+DP adversaries are `label-flipping`, `k-insertion`, `data-modification`, and
+`data-transform`. The available DP learners are `trim`, `atrim`, `irl`, and
+`orl`. There is a testing suite defined in
+`adlib/tests/learners/dp_learner_test.py`. In this file, the class
+`TestDataPoisoningLearner` is used to provide a contained testing environment
+that targets by default a ~20% poisoning percentage. To see usage, see the
+aforementioned file. If you want to run many tests of all learners with a
+specified attacker, you should use
+`adlib/tests/learners/dp_learner_many_test.py`. For example, to run 50
+`label-flipping` attacks and get detailed output of how every learner performs,
+do `python3 adlib/tests/learners/dp_learner_many_test.py 50 label-flipping`.
+This will output 50 files with the resulting tests in verbose mode.