diff --git a/src/omnipy/data/dataset.py b/src/omnipy/data/dataset.py
index 6673a380..80bf9ee4 100644
--- a/src/omnipy/data/dataset.py
+++ b/src/omnipy/data/dataset.py
@@ -15,7 +15,7 @@
 
 # from orjson import orjson
 from pydantic import Field, PrivateAttr, root_validator, ValidationError
-from pydantic.fields import Undefined
+from pydantic.fields import Undefined, UndefinedType
 from pydantic.generics import GenericModel
 from pydantic.typing import display_as_type
 from pydantic.utils import lenient_issubclass
@@ -120,9 +120,13 @@ def __class_getitem__(cls, model: ModelT) -> ModelT:
 
         return created_dataset
 
-    def __init__(self,
-                 value: Union[Dict[str, Any], Iterator[Tuple[str, Any]]] = Undefined,
-                 **input_data: Any) -> None:
+    def __init__(
+        self,
+        value: dict[str, object] | Iterator[tuple[str, object]] | UndefinedType = Undefined,
+        *,
+        data: dict[str, object] | UndefinedType = Undefined,
+        **input_data: object,
+    ) -> None:
         # TODO: Error message when forgetting parenthesis when creating Dataset should be improved.
         #       Unclear where this can be done, if anywhere? E.g.:
         #           a = Dataset[Model[int]]
@@ -136,13 +140,30 @@ def __init__(self,
         #       Dataset[Model[str]](Model[int](5)) ==  Dataset[Model[str]](data=Model[int](5))
         #       == Dataset[Model[str]](data={'__root__': Model[str]('5')})
 
+        super_kwargs = {}
+
+        assert DATA_KEY not in input_data, \
+            ('Not allowed with"data" as input_data key. Not sure how you managed this? Are you '
+             'trying to break Dataset init on purpose?')
+
         if value != Undefined:
-            input_data[DATA_KEY] = value
+            assert data == Undefined, \
+                'Not allowed to combine positional and "data" keyword argument'
+            assert len(input_data) == 0, 'Not allowed to combine positional and keyword arguments'
+            super_kwargs[DATA_KEY] = value
+
+        if data != Undefined:
+            assert len(input_data) == 0, \
+                "Not allowed to combine 'data' with other keyword arguments"
+            super_kwargs[DATA_KEY] = data
+
+        if DATA_KEY not in super_kwargs and len(input_data) > 0:
+            super_kwargs[DATA_KEY] = input_data
 
         if self.get_model_class() == ModelT:
             self._raise_no_model_exception()
 
-        GenericModel.__init__(self, **input_data)
+        GenericModel.__init__(self, **super_kwargs)
         UserDict.__init__(self, self.data)  # noqa
         if not self.__doc__:
             self._set_standard_field_description()
diff --git a/src/omnipy/data/model.py b/src/omnipy/data/model.py
index 668967c6..e10711cb 100644
--- a/src/omnipy/data/model.py
+++ b/src/omnipy/data/model.py
@@ -244,7 +244,7 @@ def __new__(cls, value: Union[Any, UndefinedType] = Undefined, **kwargs):
     def __init__(
         self,
         value: Union[Any, UndefinedType] = Undefined,
-        /,
+        *,
         __root__: Union[Any, UndefinedType] = Undefined,
         **data: Any,
     ) -> None:
@@ -263,7 +263,7 @@ def __init__(
             super_data[ROOT_KEY] = cast(RootT, data)
             num_root_vals += 1
 
-        assert num_root_vals <= 1
+        assert num_root_vals <= 1, 'Not allowed to provide root data in more than one argument'
 
         super().__init__(**super_data)
 
diff --git a/tests/data/test_dataset.py b/tests/data/test_dataset.py
index 61028ce6..9562052a 100644
--- a/tests/data/test_dataset.py
+++ b/tests/data/test_dataset.py
@@ -61,6 +61,29 @@ def test_init_with_basic_parsing():
     assert dataset_3['obj_type_2'] == '123'
     assert dataset_3['obj_type_3'] == 'True'
 
+    dataset_4 = Dataset[Model[dict[int, int]]](file_1={1: 1234, 2: 2345}, file_2={2: 2345, 3: 3456})
+
+    assert len(dataset_4) == 2
+    assert dataset_4['file_1'] == {1: 1234, 2: 2345}
+    assert dataset_4['file_2'] == {2: 2345, 3: 3456}
+
+
+def test_init_errors():
+    with pytest.raises(TypeError):
+        Dataset[Model[int]]({'file_1': 123}, {'file_2': 234})
+
+    with pytest.raises(AssertionError):
+        Dataset[Model[int]]({'file_1': 123}, data={'file_2': 234})
+
+    with pytest.raises(AssertionError):
+        Dataset[Model[int]]({'file_1': 123}, file_2=234)
+
+    with pytest.raises(AssertionError):
+        Dataset[Model[int]](data={'file_1': 123}, file_2=234)
+
+    with pytest.raises(ValidationError):
+        Dataset[Model[int]](data=123)
+
 
 def test_parsing_none_allowed():
     class NoneModel(Model[NoneType]):
diff --git a/tests/data/test_model.py b/tests/data/test_model.py
index 6475b854..7672058d 100644
--- a/tests/data/test_model.py
+++ b/tests/data/test_model.py
@@ -64,6 +64,21 @@ class StrModel(Model[str]):
     assert Model[Dict]((('a', 2), ('b', True))).to_data() == {'a': 2, 'b': True}
 
 
+def test_error_init():
+    with pytest.raises(TypeError):
+        assert Model[tuple[int, ...]](12, 2, 4).to_data() == 12
+    assert Model[tuple[int, ...]]((12, 2, 4)).to_data() == (12, 2, 4)
+
+    with pytest.raises(AssertionError):
+        Model[int](123, __root__=234)
+
+    with pytest.raises(AssertionError):
+        Model[int](123, other=234)
+
+    with pytest.raises(AssertionError):
+        Model[int](__root__=123, other=234)
+
+
 def test_load():
     model = Model[int]()
     model.from_data(12)