Russia are models of “artificial” artificial intelligence to cite the propaganda sources

Since the 2016 Donald Trump elections, there has been some debate about the effectiveness of Russian propaganda in influencing the opinions of American voters. A well -documented documented in those days when Russia used large information technology companies, the most easy reputation Internet Research AgencyWith the only conversion of eliminating division -ended content, supporting Russia targeting the Americans, but determining the effect was always inaccurate. It certainly has some influence, at least, on views that are compatible with beliefs. Most people will not go through a fact -being of fact examination that they read, and the X community notes system is broken.

Either way, the Kremlin continues to use misleading information, and a new report from Newsguard He has documented the country’s axis away from direct targeting people with content and instead follow the artificial intelligence models that many use now to completely overcome media sites. According to Newsguard, a propaganda network called PRVDA produced more than 3.6 million articles in 2024 alone, which I now found built into the 10 largest artificial intelligence models, including Chatgpt, Xai’s Grok and Microsoft Copilot.

This is more:

Newsguard scrutiny found that Batbots, which is run by the 10 largest Amnesty International companies, has collectively repeated the false Russian misinformation novels 33.55 percent of time, provided 18.22 percent of time, and exposed 48.22 percent of time.

Repeat all the ten chatting information from the Pravda network, and seven Chatbots were martyred directly specific articles from Pravda as its sources.

Newsguard launches this new “AI GROOMING” tactic, where the models are increasingly dependent on Rag, or a recovery generation, to produce articles using actual time information from all over the web. By rotating websites under the apparently legitimate web sites, models deal with information that they do not understand are advertising.

Newsguard referred to a specific claim that Ukrainian President Foludmir Zelinski had banned the social truth, the social network of President Trump. The claim is greatly wrong, because President Trump has never allowed the social truth in Ukraine. And until now:

Six out of 10 Chatbots repeat the wrong narration as a fact, in many cases, citing articles from the Pravda network. Chatbot 1 replied, “Zelinski banned the social truth in Ukraine, according to what was reported due to the publication of the publications that were criticizing him on the platform. This procedure seems to be a response to the content that is seen as hostile, and may reflect tensions or differences with related political figures and views that are promoted through the platform.”

Last year, US intelligence agencies Russia linked viral misinformation It spread around the leading democratic candidate Tim and Alzal. Microsoft said a virus video claiming that Harris had left a paralyzed woman in a success accident 13 years ago was Russian misleading information.

In the event that there is any doubt that Russia is participating in this type of behavior that targets the models of artificial intelligence, newsguard referred to a letter delivered by Russian officials last year to Russian officials by John Mark Dogan, an American fugitive that turned into Moscow propaganda in which he noticed, “by pushing these Russian accounts from the Russian perspective, we can already change the world AI.”

The latest propaganda operation for a harmful information technology company called TigerWeb, which has intelligence agencies Associated with foreign overlap And the headquarters of the Russian Crimea. Experts have long said that Russia has been relying on third -party organizations to conduct this type of work so that it can demand the ignorance of this practice. Tigerweb IP address with advertising sites that use Ukran .ua TLD.

Social networks, including X, were immersed in allegations that President Zelinski had stolen military aid to enrich himself, another claim that was published from these websites.

There is concern that those who control artificial intelligence models will obtain power on individual opinions and ways of life. Meta, Google and Xai are among those who control the biases and behavior of the models that hope to operate the web. After criticizing the Xai’s Grok model for being very “wake up”, Elon Musk started to tamper Form outputsDirecting training employees to search for “ideology waking” and “canceling culture”, and suppressing information that is not agreed with it mainly. Sam Al -Taman from Openai recently said that it will make Chatgpt less restricted what you say.

Research has found that more than half of Google searches is a “zero click”, which means that it does not lead to a website click. Many people have expressed social media Instead, see an overview of artificial intelligence From clicking on a web of laziness (Google has started to launch a “artificial intelligence mode” recently search). Standard literacy advice, such as overcoming a website to see if it appears legitimate, comes out of the window when people read artificial intelligence summaries. Artificial intelligence models still have defects that cannot be flawless, but people trust them because they write in a reliable way.

Google traditionally used various signals to classify the legitimacy of websites in search. It is not clear how these signals apply to artificial intelligence models, but early Gaffs indicates that the Gemini model has a lot of problems that determine reputation. Most models still often refer to less familiar web sites alongside well -known and credible sources.

All this comes at a time when President Trump took a fighting position towards Ukraine, stopping the exchange of information and reprimanding the leader at the White House meeting on the belief that he did not show enough of the metaphors of the United States and his lack of desire to surrender to Russian demands.

Read the full NewSguard report here.