diff --git a/services/worker/src/worker/utils.py b/services/worker/src/worker/utils.py
index 1dbc91efe..3ca0a5b36 100644
--- a/services/worker/src/worker/utils.py
+++ b/services/worker/src/worker/utils.py
@@ -8,7 +8,8 @@
 from collections.abc import Iterable
 from dataclasses import dataclass, field
 from itertools import count, islice
-from typing import Literal, Optional, TypeVar, Union, overload
+from typing import Any, Literal, Optional, TypeVar, Union, overload
+from unittest.mock import patch
 from urllib.parse import quote
 
 import PIL
@@ -38,6 +39,13 @@
 # ^ see https://pillow.readthedocs.io/en/stable/reference/Image.html#PIL.Image.MAX_IMAGE_PIXELS
 
 
+def _no_op_decode_example(self: Any, value: dict[str, Any], token_per_repo_id: Any = None) -> dict[str, Any]:  # noqa: ARG001
+    return value
+
+
+disable_video_decoding = patch("datasets.Video.decode_example", _no_op_decode_example)
+
+
 @retry(on=[ConnectionError])
 def get_rows(
     dataset: str,
@@ -65,7 +73,8 @@ def get_rows(
         raise TypeError("load_dataset should return a Dataset in normal mode")
     if column_names:
         ds = ds.select_columns(column_names)
-    rows_plus_one = list(itertools.islice(ds, rows_max_number + 1))
+    with disable_video_decoding:
+        rows_plus_one = list(itertools.islice(ds, rows_max_number + 1))
     # ^^ to be able to detect if a split has exactly ROWS_MAX_NUMBER rows
     rows = rows_plus_one[:rows_max_number]
     all_fetched = len(rows_plus_one) <= rows_max_number