diff --git a/zerver/lib/upload.py b/zerver/lib/upload.py
index 1f249b19e7..d60c3769ac 100644
--- a/zerver/lib/upload.py
+++ b/zerver/lib/upload.py
@@ -284,8 +284,6 @@ def check_upload_within_quota(realm: Realm, uploaded_file_size: int) -> None:
 def get_file_info(request: HttpRequest, user_file: File) -> Tuple[str, int, Optional[str]]:
 
     uploaded_file_name = user_file.name
-    assert isinstance(uploaded_file_name, str)
-
     content_type = request.GET.get('mimetype')
     if content_type is None:
         guessed_type = guess_type(uploaded_file_name)[0]
diff --git a/zerver/tests/test_upload.py b/zerver/tests/test_upload.py
index a247381d16..e25f247f82 100644
--- a/zerver/tests/test_upload.py
+++ b/zerver/tests/test_upload.py
@@ -39,7 +39,6 @@ from zerver.lib.actions import (
 from zerver.lib.cache import get_realm_used_upload_space_cache_key, cache_get
 from zerver.lib.create_user import copy_user_settings
 from zerver.lib.users import get_api_key
-from zerver.views.upload import upload_file_backend
 
 import urllib
 import ujson
@@ -131,43 +130,6 @@ class FileUploadTest(UploadSerializeMixin, ZulipTestCase):
         uri = result.json()["uri"]
         self.assertTrue(uri.endswith("pasted_file.png"))
 
-    def test_filename_encoding(self) -> None:
-        """
-        In Python 2, we need to encode unicode filenames (which converts them to
-        str) before they can be rendered correctly.  However, in Python 3, the
-        separate unicode type does not exist, and we don't need to perform this
-        encoding.  This test ensures that we handle filename encodings properly,
-        and does so in a way that preserves 100% test coverage for Python 3.
-        """
-
-        user_profile = self.example_user('hamlet')
-
-        mock_file = mock.Mock()
-        mock_file._get_size = mock.Mock(return_value=1024)
-
-        mock_files = mock.Mock()
-        mock_files.__len__ = mock.Mock(return_value=1)
-        mock_files.values = mock.Mock(return_value=[mock_file])
-
-        mock_request = mock.Mock()
-        mock_request.FILES = mock_files
-
-        # str filenames should not be encoded.
-        mock_filename = mock.Mock(spec=str)
-        mock_file.name = mock_filename
-        with mock.patch('zerver.views.upload.upload_message_image_from_request'):
-            result = upload_file_backend(mock_request, user_profile)
-        self.assert_json_success(result)
-        mock_filename.encode.assert_not_called()
-
-        # Non-str filenames should be encoded.
-        mock_filename = mock.Mock(spec=None)  # None is not str
-        mock_file.name = mock_filename
-        with mock.patch('zerver.views.upload.upload_message_image_from_request'):
-            result = upload_file_backend(mock_request, user_profile)
-        self.assert_json_success(result)
-        mock_filename.encode.assert_called_once_with('ascii')
-
     def test_file_too_big_failure(self) -> None:
         """
         Attempting to upload big files should fail.
diff --git a/zerver/views/upload.py b/zerver/views/upload.py
index affc06d437..59146bcdfc 100644
--- a/zerver/views/upload.py
+++ b/zerver/views/upload.py
@@ -1,5 +1,3 @@
-# -*- coding: utf-8 -*-
-
 from django.http import HttpRequest, HttpResponse, HttpResponseForbidden, \
     HttpResponseNotFound
 from django.shortcuts import redirect
@@ -73,30 +71,5 @@ def upload_file_backend(request: HttpRequest, user_profile: UserProfile) -> Http
             settings.MAX_FILE_UPLOAD_SIZE))
     check_upload_within_quota(user_profile.realm, file_size)
 
-    if not isinstance(user_file.name, str):
-        # It seems that in Python 2 unicode strings containing bytes are
-        # rendered differently than ascii strings containing same bytes.
-        #
-        # Example:
-        # >>> print('\xd3\x92')
-        # Ӓ
-        # >>> print(u'\xd3\x92')
-        # Ó
-        #
-        # This is the cause of the problem as user_file.name variable
-        # is received as a unicode which is converted into unicode
-        # strings containing bytes and is rendered incorrectly.
-        #
-        # Example:
-        # >>> import urllib.parse
-        # >>> name = u'%D0%97%D0%B4%D1%80%D0%B0%D0%B2%D0%B5%D0%B8%CC%86%D1%82%D0%B5.txt'
-        # >>> print(urllib.parse.unquote(name))
-        # Здравейте  # This is wrong
-        #
-        # >>> name = '%D0%97%D0%B4%D1%80%D0%B0%D0%B2%D0%B5%D0%B8%CC%86%D1%82%D0%B5.txt'
-        # >>> print(urllib.parse.unquote(name))
-        # Здравейте.txt  # This is correct
-        user_file.name = user_file.name.encode('ascii')
-
     uri = upload_message_image_from_request(request, user_file, user_profile)
     return json_success({'uri': uri})