diff --git a/haystack/components/extractors/named_entity_extractor.py b/haystack/components/extractors/named_entity_extractor.py
index 93fe20ece..cbb96ab46 100644
--- a/haystack/components/extractors/named_entity_extractor.py
+++ b/haystack/components/extractors/named_entity_extractor.py
@@ -5,7 +5,7 @@
 from abc import ABC, abstractmethod
 from contextlib import contextmanager
 from dataclasses import dataclass
-from enum import Enum, EnumMeta
+from enum import Enum
 from typing import Any, Dict, List, Optional, Union
 
 from haystack import ComponentError, DeserializationError, Document, component, default_from_dict, default_to_dict
@@ -21,28 +21,7 @@ with LazyImport(message="Run 'pip install spacy'") as spacy_import:
     from spacy import Language as SpacyPipeline
 
 
-class _BackendEnumMeta(EnumMeta):
-    """
-    Metaclass for fine-grained error handling of backend enums.
-    """
-
-    def __call__(cls, value, names=None, *, module=None, qualname=None, type=None, start=1):  # noqa: A002
-        if names is None:
-            try:
-                return EnumMeta.__call__(cls, value, names, module=module, qualname=qualname, type=type, start=start)
-            except ValueError:
-                supported_backends = ", ".join(sorted(v.value for v in cls))  # pylint: disable=not-an-iterable
-                raise ComponentError(
-                    f"Invalid backend `{value}` for named entity extractor. "
-                    f"Supported backends: {supported_backends}"
-                )
-        else:
-            return EnumMeta.__call__(  # pylint: disable=too-many-function-args
-                cls, value, names, module, qualname, type, start
-            )
-
-
-class NamedEntityExtractorBackend(Enum, metaclass=_BackendEnumMeta):
+class NamedEntityExtractorBackend(Enum):
     """
     NLP backend to use for Named Entity Recognition.
     """
@@ -53,6 +32,24 @@ class NamedEntityExtractorBackend(Enum, metaclass=_BackendEnumMeta):
     #: Uses a spaCy model and pipeline.
     SPACY = "spacy"
 
+    def __str__(self):
+        return self.value
+
+    @staticmethod
+    def from_str(string: str) -> "NamedEntityExtractorBackend":
+        """
+        Convert a string to a NamedEntityExtractorBackend enum.
+        """
+        enum_map = {e.value: e for e in NamedEntityExtractorBackend}
+        mode = enum_map.get(string)
+        if mode is None:
+            msg = (
+                f"Invalid backend '{string}' for named entity extractor. "
+                f"Supported backends are: {list(enum_map.keys())}"
+            )
+            raise ComponentError(msg)
+        return mode
+
 
 @dataclass
 class NamedEntityAnnotation:
@@ -134,7 +131,7 @@ class NamedEntityExtractor:
         """
 
         if isinstance(backend, str):
-            backend = NamedEntityExtractorBackend(backend)
+            backend = NamedEntityExtractorBackend.from_str(backend)
 
         self._backend: _NerBackend
         device = ComponentDevice.resolve_device(device)
diff --git a/releasenotes/notes/fix-named-entity-entity-extractor-backend-enum-2e64e25f8d7f1b08.yaml b/releasenotes/notes/fix-named-entity-entity-extractor-backend-enum-2e64e25f8d7f1b08.yaml
new file mode 100644
index 000000000..0df95101c
--- /dev/null
+++ b/releasenotes/notes/fix-named-entity-entity-extractor-backend-enum-2e64e25f8d7f1b08.yaml
@@ -0,0 +1,4 @@
+---
+fixes:
+  - |
+    Fix `NamedEntityExtractor` crashing in Python 3.12 if constructed using a string backend argument.