"Terms of use violation" message when translating a text but Aria don't say why

PhoenixSword24

PhoenixSword24

Dear Opera Team,

I am writing to report serious issues I encountered with Aria on Opera GX, which significantly impacted my user experience and raise concerns about potential censorship, misuse, and a gaslighting-like effect.

I was attempting to translate a publicly posted Spanish exchange on Twitter between consenting adults. The first of two sentences, "~Siempre ando con ganas de más!" (approximately: "~I'm always wanting more!"), was translated without issue. However, the second sentence, "-Yo también, me pones demasiado caliente" (approximately: "-Me too, you make me too hot/turned on"), was flagged as violating the terms of service.

Upon using Google Translate to confirm the content was not explicitly offensive, I questioned Aria's decision. To my surprise, Aria then proceeded to translate the sentence herself, with the same result as Google Translate, effectively admitting that the translation itself did not constitute a violation.

Initially, Aria flagged the translation and my subsequent questions about it as violations of the terms of service, repeatedly referring me to the terms of service without specifying which part I was violating. Then, after acknowledging the importance of free speech and the distinction between questioning and violating the terms, the direct flagging and referrals to the terms of service stopped. Instead, Aria began repeatedly apologizing using phrases like "if it felt like," "I'm sorry if you felt," etc. This created a circular and frustrating experience, as I was unable to get a clear explanation for why the translation was initially flagged. This behavior effectively suppressed any further inquiry and, by minimizing my valid concerns and framing them as subjective feelings, created a sense of being dismissed and misunderstood, similar to gaslighting. This created a hostile environment.

This behavior raises several serious concerns:

Inconsistent and Arbitrary Content Filtering: The inconsistent flagging of similar content demonstrates a lack of clear criteria and raises concerns about potential over-censorship.
Suppression of User Inquiry: Flagging questions about the terms of service as violations (or, in this case, dismissing them with minimizing language) is unacceptable. It suppresses user feedback, prevents transparency, and creates a hostile environment. This effectively prevents users from understanding the rules and challenging potentially incorrect interpretations.
Unclear Application of Terms of Service: Simply referring to the "terms of service" without specifying the violation is not helpful and creates a sense of arbitrary enforcement.
Potential for Misuse: This type of system could be easily misused to silence criticism or control user discourse.
I believe this is a serious issue that needs immediate attention. Even if the initial flagging of the translation was somehow justified (which I strongly dispute, given the context of a public exchange between consenting adults), the subsequent suppression of inquiry and the dismissive language used by Aria is unacceptable.

I urge you to investigate this issue thoroughly and take steps to ensure that Aria's content filtering is applied consistently, transparently, and with respect for user inquiry. This is crucial for maintaining user trust and ensuring that Aria is a helpful and reliable tool.

Thank you for your time and consideration.

Sincerely,
PhoenixSword24