Rahmat Wibowo published Meta AI's 8.5 rating of his somasi to Hasbi A Lubis

He published a page under Hasbi A Lubis's name about a Meta AI reply, addressed to Hasbi's account, that rated his legal notice 099 '8.5 out of 10', while writing that he does 'not treat it as an authority' and may be 'giving the review too much room by writing about it'.

Original post ↗

Rahmat Wibowo published Meta AI's 8.5 rating of his somasi to Hasbi A Lubis

Commentary

What he posted

He published it on InfraLoka’s blog on 14 November 2026 under “Professional Blacklist”, titled “A Meta AI Review of Somasi 099 to Hasbi A Lubis: What an AI Comment Is and Is Not”. He writes that under his thread that day “an account called meta.ai posted a rating of my letter, addressed to @hasbilubistrilife”, and that “I did not write it and I do not treat it as an authority.” He lists what it said: it gave the letter “a rating of 8.5 out of 10 as a self-written letter by a non-lawyer”, said “the usual letters on social media get 3 to 4”, praised his choice of articles and “the open admission of weak points, such as unconfirmed identity”, and said “the 11-page public version is long for one comment” and that “a lawyer is still needed” if he goes further. Under “Why it counts for little” he writes: “A flattering review is easy to quote. I do not want to lean on one.” He says he publishes it “only because it is part of the public thread”, and under “Where I am probably wrong” adds: “I may be giving the review too much room by writing about it.”

Context

The review concerns legal notice 099, whose terms he published the same day (ev-20261114-010). That morning he told Hasbi A Lubis “I am Parashurama” (ev-20261114-008) and reported Hasbi to Vopak’s whistleblowing channel (ev-20261114-005). He has asked Meta AI to grade his somasi before, pushing it to raise its score for his somasi to Kukuh Setiawan (ev-20261017-004).

Observations

A flattering review given a page of its own

He writes that a flattering review “is easy to quote” and that he does not want to lean on one. He still gives it a full page under Hasbi A Lubis’s name, with the 8.5 rating in the text. Publishing the page is itself leaning on it.

An AI rating aimed at the person he targets

The rating was addressed to Hasbi A Lubis’s account, under his own thread. He writes that he does not know how Hasbi saw it, or whether it was “fair to him”. A public score of a legal notice, directed at its recipient, adds pressure without adding any fact.

The review’s own caveats point to overreach

Meta AI said the 11-page public version “is long for one comment” and flagged the unconfirmed identity. He repeats both as weak points he “already wrote down”. The page records them and changes nothing about the letter or its deadlines.

Transcript

Home / Professional Blacklist / A Meta AI Review of Somasi 099 to Hasbi A Lubis: What an AI Comment Is and Is Not Professional Blacklist: personal dispute documentation Professional Blacklist Hasbi A Lubis Somasi Right Of Reply A Meta AI Review of Somasi 099 to Hasbi A Lubis: What an AI Comment Is and Is Not Rahmat Wibowo · November 14, 2026 Under my thread on 14 November 2026, an account called meta.ai posted a rating of my letter, addressed to @hasbilubistrilife. I did not write it and I do not treat it as an authority. This post explains why. This article does not claim that Hasbi A Lubis asked for, read or reacted to the review. The screenshots do not show who started it, and I make no claim about that. This is my account and my opinion, not a finding by any court. For context see the main record. What the reply said In two screenshots (taken at about 08.24 WIB), the account meta.ai: gave my letter, Somasi No. 099/SOMASI/RW/XI/2026, a rating of 8.5 out of 10 as a self-written letter by a non-lawyer, and said the usual letters on social media get 3 to 4; praised the use of the 2023 KUHP (Pasal 436, 441 and 440) over the old ITE article, and the choice not to rely on Pasal 433 or 434; praised the table of elements and the open admission of weak points, such as unconfirmed identity; said the 11-page public version is long for one comment and that a lawyer is still needed to check the evidence hashes if I go further; said it was not legal advice. Why it counts for little It is software output. It can be wrong, and it has no duty to me or to anyone else. It reads my letter and stops there. It cannot check the facts behind it. A good rating is not a legal finding. No court, police unit or lawyer has endorsed the letter. A flattering review is easy to quote. I do not want to lean on one. I publish it only because it is part of the public thread and because it repeats the weak points I already wrote down: the unconfirmed identity, the small weight of one comment, and the need for a lawyer if this goes further. What I cannot say That the review is correct. That it was fair to him. It was addressed to him, and I do not know how he saw it. That the letter is strong. Only a court can say. Where I am probably wrong I may be giving the review too much room by writing about it. A fair reader could say a rating by an AI belongs nowhere in a record. I include it for completeness and nothing more. Right of reply and correction policy Hasbi A Lubis is invited to send a response, correction or added context. A substantive response may be published alongside this article, subject to reasonable editing for privacy, safety and relevance. If reliable evidence shows that any factual statement here is wrong, I will correct it promptly and note the change. Published by Rahmat Wibowo, InfraLoka — infraloka.co.id #ProfessionalBlacklist #HasbiALubis #Somasi #RightOfReply