Ill have to dig into that over the next day or so but a quick search provided me with this information
The accuracy of AI stylometry analysis can vary depending on the complexity of the situation, but generally reaches high levels, with studies reporting accuracies often exceeding 80% and sometimes reaching near 90% when identifying authorship in controlled scenarios, especially when comparing between only a few potential authors; however, accuracy can significantly decrease when dealing with a large pool of possible authors or when authors intentionally try to obfuscate their writing style
Is it your expert opinion without a doubt that date of December 9th could not have been altered after the fact? Is there no chance things may not be as they seem?
Are there forensic web data analysis tools that the general public generally does not have access to that could be implemented to further investigate this domain?