diff --git a/haystack/schema.py b/haystack/schema.py
index 1b21daa94..7078afe08 100644
--- a/haystack/schema.py
+++ b/haystack/schema.py
@@ -1,5 +1,6 @@
 from __future__ import annotations
 import csv
+import hashlib
 
 import typing
 from typing import Any, Optional, Dict, List, Union
@@ -635,7 +636,7 @@ class MultiLabel:
 
         self.query = self._aggregate_labels(key="query", must_be_single_value=True)[0]
         self.filters = self._aggregate_labels(key="filters", must_be_single_value=True)[0]
-        self.id = hash((self.query, json.dumps(self.filters, sort_keys=True).encode()))
+        self.id = hashlib.md5((self.query + json.dumps(self.filters, sort_keys=True)).encode()).hexdigest()
 
         # Currently no_answer is only true if all labels are "no_answers", we could later introduce a param here to let
         # users decided which aggregation logic they want
diff --git a/test/others/test_schema.py b/test/others/test_schema.py
index d1af0d02d..f83f91477 100644
--- a/test/others/test_schema.py
+++ b/test/others/test_schema.py
@@ -361,6 +361,46 @@ def test_multilabel_preserve_order_w_duplicates():
         assert multilabel.labels[i].id == str(i)
 
 
+def test_multilabel_id():
+    query1 = "question 1"
+    query2 = "question 2"
+    document1 = Document(content="something", id="1")
+    answer1 = Answer(answer="answer 1")
+    filter1 = {"name": ["name 1"]}
+    filter2 = {"name": ["name 1"], "author": ["author 1"]}
+    label1 = Label(
+        query=query1,
+        document=document1,
+        is_correct_answer=True,
+        is_correct_document=True,
+        origin="gold-label",
+        answer=answer1,
+        filters=filter1,
+    )
+    label2 = Label(
+        query=query2,
+        document=document1,
+        is_correct_answer=True,
+        is_correct_document=True,
+        origin="gold-label",
+        answer=answer1,
+        filters=filter2,
+    )
+    label3 = Label(
+        query=query1,
+        document=document1,
+        is_correct_answer=True,
+        is_correct_document=True,
+        origin="gold-label",
+        answer=answer1,
+        filters=filter2,
+    )
+
+    assert MultiLabel(labels=[label1]).id == "33a3e58e13b16e9d6ec682ffe59ccc89"
+    assert MultiLabel(labels=[label2]).id == "1b3ad38b629db7b0e869373b01bc32b1"
+    assert MultiLabel(labels=[label3]).id == "531445fa3bdf98b8598a3bea032bd605"
+
+
 def test_serialize_speech_document():
     speech_doc = SpeechDocument(
         id=12345,