Reading restriction on Twitter: "Musk stepped on a rattlesnake"

02.07.2023

Twitter owner Elon Musk has justified the massive restrictions on functions of the service with the comprehensive evaluation of data by third parties. One observer confirmed that Twitter is the most important platform for the "censorship industry" to train its AI models.

The short message service Twitter severely limited the functions of the platform on Saturday. Verified accounts, which pay eleven euros per month, could only read 6,000 posts per day. Unverified accounts were limited to 600 posts per day, and new, unverified accounts were limited to 300 posts per day. Since scrolling over a tweet is already considered reading, the limit can be reached very quickly, especially for unverified accounts. Using Twitter without an account is practically no longer possible.

The owner of the platform, Elon Musk, justified the drastic restrictions in a tweet on Saturday with the "extreme extent" that the evaluation of tapped data (web scraping) and system manipulation have reached:

"Nearly every company involved in artificial intelligence, from start-ups to some of the largest companies in the world, has tapped huge amounts of data."

Several hundred organizations have tapped Twitter data so "extremely aggressively" that "the user experience has been compromised," Musk said. The restrictions are temporary and will be lifted or relaxed. Verified accounts will soon be able to read up to 8,000 posts again, unverified 800 and new, verified 400. Later, Musk increased the figures to 10,000, 1,000 and 500.

The reactions of Twitter users were marked by a lack of understanding, such as why Twitter does not use captcha or other web scraping protections that other platforms use without any problems.

On Sunday morning, Twitter Daily News wrote that the short message service had to put emergency servers online in order to handle the capacity for data readout. Twitter has limited capacity as a company and cannot cope with these illegal activities. The anger of the users should be directed at the illegal actions of the AI companies that carried out the aggressive data readout.

Optimization of AI censorship models

Mike Benz, executive director of the Foundation for Freedom Online and an information technology specialist, took to Twitter on Saturday to make some reasonable assumptions about the background to Twitter's restrictions.

Benz pointed out that since 2016 there have been constantly evolving AI censorship models based primarily on the mass evaluation of Twitter data. This is due to the fact that, unlike Facebook, for example, Twitter is not a "fenced garden" where you may have to be friends with other users to see the content of their pages.

Unlike on YouTube, on Twitter every user is also a creator of content as soon as he presses the retweet button. However, the CIA, the U.S. State Department, the U.S. Department of Defense, charitable foundations, and university "research" centers abused the platform's openness to use their AI models to map the emergence of narratives in real time. Benz:

"To build a Death Star of media censorship, they need to be able to pick up hundreds of millions of tweets."

Business reasons or fight against censorship?

According to Benz, how Musk is currently acting is interesting because, on the one hand, he is restricting the openness of the Internet itself by taking measures against data readout. On the other hand, however, it prevents the further optimization of a "censorship Death Star".

Benz could not rule out the possibility that Musk might end up trying to increase the number of paying, verified accounts only for business reasons. After all, these are not so badly affected by the reading restrictions. But he is sure that this week there will be hundreds of censorship agents who are housed in the research centers of the universities and are now "howling at the moon" because they can no longer do their job.