Ask an AI for tests and you will get tests, and they will usually pass. The question is whether they would fail if the code were wrong. A test that cannot fail is decoration.
The common shapes: an assertion so loose any value passes ("is defined", "is not null"); a try block that swallows the error the test was meant to expect; a test that mocks the very function it claims to test and then checks the mock; the one important test marked skip.
The quick way to find out is to break the code on purpose: change a + to a -, return early, delete a check. Run the tests. If they stay green, they were not testing that. Mutation-testing tools (Stryker for JavaScript, Infection for PHP, mutmut for Python) do this automatically.
When you ask for tests, ask for the failing case first: "write the test, show me it failing, then write the code."