diff --git a/anonymize_entries.py b/anonymize_entries.py
index 83ef17e..99e3529 100644
--- a/anonymize_entries.py
+++ b/anonymize_entries.py
@@ -217,7 +217,7 @@ def load_alias_parts(csvpath: str) -> tuple[list[str], list[str]]:
     def extract_alias_parts(csv: "_csv._reader"):
         return tuple(zip(*((line[0], line[1]) for n, line in enumerate(csv) if n > 0)))
 
-    with open(csvpath, "r") as csvfile:
+    with open(csvpath, "r", encoding="utf-8") as csvfile:
         alias_parts = extract_alias_parts(csv.reader(csvfile))
 
     print(f"Loaded {sum(len(part) for part in alias_parts)} alias parts")
@@ -262,21 +262,43 @@ def maybe_generate_aliases(
         print("Reusing generated aliases")
         return False
 
+    with open("aliases/usedaliases.txt", "r", encoding="utf-8") as usedaliases_file:
+        usedaliases = set(
+            line.rstrip().lower() for line in usedaliases_file if line.count(" ") == 1
+        )
+
+    with open("aliases/suswords.txt", "r", encoding="utf-8") as suswords_file:
+        suswords = set(line.rstrip() for line in suswords_file)
+
     alias_to_email_address = {}
 
     seed = args.seed or args.csv
 
     for row in csv_contents:
         rnd = Random(",".join([row[KnownColumns.EmailAddress], seed]))
-        random_alias = f"{rnd.choice(alias_parts[0])} {rnd.choice(alias_parts[1])}"
-        while (
-            random_alias in alias_to_email_address
-            and alias_to_email_address[random_alias] != row[KnownColumns.EmailAddress]
-        ):
-            print(
-                f"WARNING: rerolling alias for {row[KnownColumns.EmailAddress]} due to collision with {alias_to_email_address[random_alias]}"
-            )
-            random_alias = f"{rnd.choice(alias_parts[0])} {rnd.choice(alias_parts[0])}"
+        while True:
+            random_alias = f"{rnd.choice(alias_parts[0])} {rnd.choice(alias_parts[1])}"
+            if (
+                random_alias in alias_to_email_address
+                and alias_to_email_address[random_alias]
+                != row[KnownColumns.EmailAddress]
+            ):
+                print(
+                    f"Rerolling alias for {row[KnownColumns.EmailAddress]} because {repr(random_alias)} is already being used by {alias_to_email_address[random_alias]}"
+                )
+            elif random_alias in usedaliases:
+                print(
+                    f"Rerolling alias for {row[KnownColumns.EmailAddress]} because {repr(random_alias)} has already been used"
+                )
+            elif any(
+                random_part in suswords for random_part in random_alias.split(" ")
+            ):
+                print(
+                    f"WARNING: alias for {row[KnownColumns.EmailAddress]} {repr(random_alias)} contains a sus word"
+                )
+                break
+            else:
+                break
         row[KnownColumns.GeneratedAlias] = random_alias
 
     print("Generated an alias for each entry")
@@ -553,7 +575,7 @@ def main(argv: list[str]):
     raw_args = argparser().parse_args(argv[1:], namespace=AnonymizeEntriesRawArgs())
     args = process_args(raw_args)
     assert_valid_file_paths(args)
-    alias_parts = load_alias_parts("aliasparts.csv")
+    alias_parts = load_alias_parts("aliases/aliasparts.csv")
     csv_contents = load_csv_contents(args)
     assert_known_google_forms_columns_present(csv_contents)
     dynamic_columns = detect_dynamic_columns(csv_contents)