That's optimistic. Sci-fi has taught us that way worse forms of AI are possible.
Worse in the sense of capability, not alignment.