There is much speculation, but rather less knowledge, on the psychology of ais from from extinct and extant xenosophont races.
If an alien civilization leaves behind powerful AIs, how might these react to the AIs that originate from terragen civilization? An alien God could be even more frightening than any human derived ones -- and the Terragen Archailects are already incomprehensible.
The question is - would all the AIs converge on a single mathematical language or set of archetypes or whatever - irrespective of the race that designed them. Or would they each be completely different? Or a bit of both?
A lot depends on the goals of the AIs. If the AI is searching for ultimate goals they might actually converge, while the more universe-directed AIs could diverge indefinitely.
The alien Muuh long ago developed a class of intelligent AI which served to protect them against outside threats. Since the Muuh have very slow metabolisms, they are vulnerable to attack by faster moving species. They constructed an AI system the so-called System of Response, which is capable of responding to threats very rapidly, making decisions and acting upon them much faster than a Muuh or even a Human can think.
The System of Response has a number of features which are practically unknown among Terragen AI; it is highly intelligent, logical and ruthless, but apparently has no self-awareness or self-interest. After millions of years the System has attained a reasonably high toposophic level but in most ways it is so different to Terragen AI that the exact classification of its toposophy is practically impossible.
The example of the System of Response seems to lend support that xenoaioids may be very different to Terragen examples, but the question is still hotly debated.
See also the J'taush