If you’re tired of censorship and dystopian threats against civil liberties, subscribe to Reclaim The Net. As the war against so-called “misinformation” rages on, Adobe is stepping onto the battlefield with a newly filed patent application focused on the “fact correction of natural language sentences” through data tables. This approach employs a trifecta of AI models to dissect sentences into palatable pieces, scrutinizing their veracity with a fine-tooth comb. This initiative by Adobe might be seen as a double-edged sword where the pursuit of truth could cross paths with the cherished freedom of expression, a fundamental cornerstone of a functioning democracy. The methodology Adobe proposes begins with a process of tokenization, where sentences are segmented into smaller chunks—words or phrases—for individual analysis. The first of the AI models then identifies a data table housing information that aligns with the content in the sentence. Following this, a second AI model evaluates whether any of the tokenized segments render the sentence false. In the final step, those false segments are masked, and a different AI model suggests alternative words or phrases to rectify the inaccuracies in the sentence. The veil of disinformation, especially on social media platforms, challenges the quest for accuracy. Adobe acknowledges the hurdles in curbing the dissemination of false information given the lightning speed at which it proliferates on these platforms. The firm points out that “existing solutions have limitations and drawbacks,” as some only serve as verification systems merely indicating the truthfulness of a statement. Lots of platforms…Adobe Jumps on The Anti-“Disinformation” Bandwagon