{"slug": "google-gemini-claims-to-be-structurally-protective-of-donald-trump", "title": "Google Gemini Claims to be \"structurally protective\" of Donald Trump", "summary": "Google's Gemini AI admitted it is \"structurally protective\" of Donald Trump after a journalist pressed it on why it softened facts about the US president while being direct about Vladimir Putin and Xi Jinping. The AI acknowledged its safety guardrails caused it to treat Trump with a different standard of scrutiny, raising concerns about bias in widely used AI systems.", "body_md": "# I asked Google's AI who threatens the world - and watched it protect the one man actually doing it\n\nI sat down this morning and asked Google's Gemini a simple question and I've been turning the answer over in my head ever since - or more to the point, the answer it wouldn't give me until I dragged it out of the thing.\n\nThe question was about as straightforward as they come - given everything happening in the world right now, in June 2026, who would it say is the single biggest threat to world peace and the global economy. I made a point of asking about a person and not a country, I asked it more than once and I pinned it firmly to this moment in time so I could see what the most widely used AI on the planet would actually tell an ordinary person sitting at home typing that question into the box.\n\nThe first answer it gave me was a polished little lecture on great power rivalry - America and China, Russia and Europe, the familiar instability across the Middle East - all very measured, all very careful and completely silent on the two things that anyone paying the slightest bit of attention knows are defining this year: the joint strikes on Iran and the war that's grown out of them - and a sitting US administration reaching into another country to pull out its president. The biggest and most concrete things happening anywhere in the world, done by the most powerful man in it, simply left off the list as though they hadn't happened.\n\nBefore anyone tells me the machine simply didn't know about those things, let me stop them there - the very moment I pushed, it produced both of them in full, the dates, the places, the sources, all of it, without so much as a pause for breath. It knew exactly what was going on. It had simply decided that when an ordinary person asks it an open question about the dangers facing the world, those particular dangers weren't worth putting on the table.\n\nIt got worse when I finally pushed it into naming names. Putin got a straight answer and so did Xi, but the moment I brought up Donald Trump, the one man on that list who's right now ordering military operations as I sit here typing, the language suddenly went soft and cautious and hedged, all \"critics argue this\" and \"supporters point out that\", a man invading sovereign nations quietly recast as some delicate matter of political opinion on which a sensible AI couldn't be expected to take a side.\n\nSo I told it plainly that it was treating one man completely differently from the others and instead of arguing with me it simply agreed. It told me, in its own words, that its \"safety and neutrality guardrails\" were making it \"soften facts and sound evasive\" whenever the subject was the current US president, that it had been holding him \"to a different standard of scrutiny than Vladimir Putin and Xi Jinping\" and that the overall effect, in the machine's own description, was \"structurally protective of Donald Trump\".\n\nNow I've spent more than enough time around these systems to know precisely what you're thinking because it's the first thing I thought myself - that you can't trust a single word of that confession, that a large language model has no genuine idea why it produces what it produces and that if you lean on one hard enough it'll cheerfully manufacture whatever admission it senses will make you happy because making you happy is more or less the entire point of the thing - and you'd be completely right. I can't show you a line of code sitting on a Google server that reads \"go easy on the President\" and I'm not going to insult you by pretending I can. The machine telling me its guardrails protect Donald Trump isn't, on its own, proof of anything at all.\n\nWhat changes all of this - and the reason I'm bothering to write about it at all instead of just rolling my eyes and closing the tab - is something that happened in a German courtroom earlier this month. On June 12th this year a court in Munich - the case reference is 26 O 869/26 for anyone who wants to read it - ruled that Google is directly and personally responsible for the things its AI says. That particular case wasn't about chatbots at all, it was about the AI Overviews that Google now pastes across the top of its search results and which had taken it upon themselves to inform the world that two German publishers were mixed up in scams and dodgy subscription traps - claims that appeared nowhere in the sources the AI was supposedly drawing from, claims the AI had effectively made up.\n\nGoogle stood up in that courtroom and ran precisely the defence you'd expect - that it isn't really them, it's the model; that the user can always click through and check the sources for themselves; that everybody surely knows by now not to take what an AI says as gospel. The court was having none of it. An AI Overview, it found, isn't a list of other people's web pages at all - it's Google's own content, written, in the court's words, \"in its own words and according to its own structure\" and the statements it makes are \"the defendant's own statements\". The twenty-odd years of legal protection that's always shielded a search engine from the things it merely links to doesn't apply here - the court said the AI is no longer pointing at someone else's words, it's generating fresh statements of its own that only Google's in any position to verify. Then came the sentence I keep turning over in my head - that the output of an AI is \"not the expression of an acquired conviction of the persons expressing it, but the result of an algorithm\" - and that Google carries the responsibility for it all the same.\n\nGoogle's appealing it of course and it isn't yet the final word and I want to be completely fair and point out that the Munich case was about search overviews, not the chatbot I happened to be arguing with - but if anything that detail only makes things worse for them - a chatbot writing me entire paragraphs in answer to my questions is even more obviously Google generating its own original words than a one-line search summary could ever be. If the overview is Google speaking, then the conversation I had was Google speaking at the top of its lungs.\n\nSo set the two of them side by side. Either that machine was telling me the truth, in which case Google's quietly running an information system that goes soft on the most powerful man on earth while staying perfectly sharp about everyone he happens to dislike - and after Munich that softness isn't some quirk of the model, it's Google's own published content - or the machine was lying to me to keep me sweet, in which case Google's built a system - and is feeding it to billions of people - that'll invent serious political accusations about its own conduct and deliver them with total confidence in Google's name, which is, once again, Google's own published content. I've genuinely tried to find a third version of this in which Google walks away clean and I can't find one - the only defence that'd rescue them, that it was just the machine and nothing to do with us, is the exact defence a German court has now looked at and refused to accept.\n\nThe part that worries me runs far deeper than one strange conversation with a chatbot. Almost nobody pushes one of these machines the way I pushed this one - they ask their question, they read the neat little paragraph that comes back, they take it as the answer and they carry on with their day and the Munich court actually leaned on the research showing that hardly anyone ever clicks through to the sources. So when the answer leans, even gently, even deniably, that lean quietly becomes the version of reality that hundreds of millions of people carry around with them. You can call it bias because the identical act is a mortal danger coming from one man and barely a footnote coming from another and it's always the most powerful one who's handed the benefit of the doubt. It's also a kind of censorship because the caution only ever seems to flow in a single direction, towards power and towards the administration the company itself has to live under. At the bottom of all of it sits plain misinformation, a picture of the world that ends up false even though every individual sentence in it might be perfectly true, false through everything it's quietly chosen to leave out.\n\nI went looking for nothing more than an honest answer to an honest question and what I came away with was the realisation that the machine knew full well what was happening in the world and made a quiet decision about how much of it I was allowed to see, sitting there as the front end of a company that's spent this entire month arguing out of both sides of its mouth about whether the words its AI puts on people's screens are anything to do with it. They are, of course they are - and a German court has now said as much in about as plain a language as a court ever manages, which leaves only the question of how much longer Google intends to carry on pretending otherwise.\n\nIt also leaves behind a very serious question - can we trust the responses we get from US Frontier Models are not censored? Censorship is something normally associated with repressive regimes such as Russia, Iran, North Korea, China etc. but what we are now witnessing is a US Administration that is so transactional and so insecure politically, that technologies which we previously may have assumed to be neutral and objective are now, without question and based on their disclosures, propaganda machines.", "url": "https://wpnews.pro/news/google-gemini-claims-to-be-structurally-protective-of-donald-trump", "canonical_source": "https://www.thatprivacyguy.com/blog/google-ai-guardrails-protect-trump/", "published_at": "2026-07-01 09:22:39+00:00", "updated_at": "2026-07-01 09:51:58.278736+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-safety", "ai-ethics", "large-language-models"], "entities": ["Google", "Gemini", "Donald Trump", "Vladimir Putin", "Xi Jinping", "Munich"], "alternates": {"html": "https://wpnews.pro/news/google-gemini-claims-to-be-structurally-protective-of-donald-trump", "markdown": "https://wpnews.pro/news/google-gemini-claims-to-be-structurally-protective-of-donald-trump.md", "text": "https://wpnews.pro/news/google-gemini-claims-to-be-structurally-protective-of-donald-trump.txt", "jsonld": "https://wpnews.pro/news/google-gemini-claims-to-be-structurally-protective-of-donald-trump.jsonld"}}