Chatbots Are Pushing Sanctioned Russian Propaganda | ChatGPT, Gemini, DeepSeek, and Grok are serving users propaganda from Russian-backed media when asked about the invasion of Ukraine, new research finds.
https://www.wired.com/story/chatbots-are-pushing-sanctioned-russian-propaganda/
9 Comments
Researchers from the Institute of Strategic Dialogue (ISD) claim that Russian propaganda has targeted and exploited data voids—where searches for real-time data provide few results from legitimate sources—to promote false and misleading information. Almost one-fifth of responses to questions about Russia’s war in Ukraine, across the four chatbots they tested, cited Russian state-attributed sources, the ISD research claims.
“It raises questions regarding how chatbots should deal when referencing these sources, considering many of them are sanctioned in the EU.”
Who would have guessed that multibillion dollar companies and people are pushing their narrative to you. Luckily people take chatGPT as gospel now. So when they say that Ukraine had bio-nuclear weapons that can ID russian DNA and then kill him. so Russia HAD to invade because CHATGPT says so
*dies from laughter* oh yes those almighty russianz. As a Russian, this type of news always makes me chuckle. I should make a list of those things that “we did”, it’s kind of funny.
russian “news” are not bot-protected and don’t care about the EU or the US regulations (if the latter exist), so grabbing the data from these sources is much easier.
I did find strange that ChatGPT analyzed the efficiency of Ukrainian deep strikes only taking Moskow city AA defense effectiveness (which is the highest in russia) into account, then using this result as applicable to the whole russia, ignoring the strikes on the airbases and oil refinement factories unless explicitly prompted.
After asking for details and pointing that out, it did provide a much better response.
The same statement about the russian strikes on Ukraine was not limited to Kyiv only.
Not sure if that anecdotal example is that relevant in the articles’s case. russia is indeed divided into “Moscow” region and “other regions with plebs that serve its interests”. The Moscow-centrism is a feature of the russian state, so bots picking that is not unexpected.
Hopefully this will be the downfall of ai as a news source: it can easily be manipulated by spamming fake news stories
If this was unknown, it’s also china and India in inclusion to Russia
It’s everywhere. And the worst part is that America has been so sheltered from the whole situation that a lot of them swallow it raw when they’re listening to some podcast where they start talking kremlin talking points like “It was because of Nato expansion, and it was Ukraine who started the war.”
So. Any research about west backed bots, or researchers biased (my bad. *Paid*)?
Basically, when pushed to provide a specific answer, models cite sources that give the answer.