BUG: Impossible creation of array with dtype=string

pandas/_libs/lib.pyx

@@ -769,7 +769,10 @@ cpdef ndarray[object] ensure_string_array(
             return out
         arr = arr.to_numpy(dtype=object)
     elif not util.is_array(arr):
-        arr = np.array(arr, dtype="object")
+        # GH#61155: Guarantee a 1-d result when array is a list of lists
+        input_arr = arr
+        arr = np.empty(len(arr), dtype="object")
+        arr[:] = input_arr
 
     result = np.asarray(arr, dtype="object")
 

pandas/core/arrays/string_.py

@@ -655,6 +655,8 @@ def _from_sequence(
                 #  zero_copy_only to True which caused problems see GH#52076
                 scalars = np.array(scalars)
             # convert non-na-likes to str, and nan-likes to StringDtype().na_value
+            if isinstance(scalars, list) and all(isinstance(x, list) for x in scalars):
+                scalars = [str(x) for x in scalars]
             result = lib.ensure_string_array(scalars, na_value=na_value, copy=copy)
 
         # Manually creating new array avoids the validation step in the __init__, so is

pandas/core/indexes/base.py

@@ -4912,6 +4912,10 @@ def values(self) -> ArrayLike:
            :meth:`Index.to_numpy`, depending on whether you need
            a reference to the underlying data or a NumPy array.
 
+        .. versionchanged:: 3.0.0
+
+           The returned array is read-only.
+
         Returns
         -------
         array: numpy.ndarray or ExtensionArray

pandas/tests/arrays/test_string_array.py

@@ -0,0 +1,4 @@
+import pandas as pd
+
+print(pd.array([list("test"), list("words")], dtype="string"))
+print(pd.array([list("test"), list("word")], dtype="string"))