When the official <\/span>White House X account posted<\/a> an image depicting activist Nekima Levy Armstrong in tears during her arrest, there were telltale signs that the image had been altered.<\/p>\n Less than an hour before, Homeland Security Secretary Kristi Noem had posted<\/a> a photo of the exact same scene, but in Noem\u2019s version Levy Armstrong appeared composed, not crying in the least.<\/p>\n Seeking to determine if the White House version of the photo had been altered using artificial intelligence tools, we turned to Google\u2019s SynthID<\/a> \u2013 a detection mechanism that Google claims is able to discern whether an image or video was generated using Google\u2019s own AI. We followed Google\u2019s instructions and used its AI chatbot, Gemini, to see if the image contained SynthID forensic markers.<\/p>\n The results were clear: The White House image had been manipulated with Google\u2019s AI. We published a story about it<\/a>.<\/p>\n After posting the article, however, subsequent attempts to use Gemini to authenticate the image with SynthID produced different outcomes.<\/p>\n In our second test, Gemini concluded that the image of Levy Armstrong crying was actually authentic. (The White House doesn\u2019t even dispute that the image was doctored. In response to questions about its X post, a spokesperson said, \u201cThe memes will continue.\u201d)<\/p>\n In our third test, SynthID determined that the image was not made with Google\u2019s AI, directly contradicting its first response.<\/p>\n At a time when AI-manipulated photos and videos are growing inescapable, these inconsistent responses raise serious questions about SynthID\u2019s reliability to tell fact from fiction.<\/p>\n Google describes<\/a> SynthID as a digital watermarking system. It embeds invisible markers into AI-generated images, audio, text or video created using Google\u2019s tools, which it can then detect \u2013 proving whether a piece of online content is authentic.<\/p>\n \u201cThe watermarks are embedded across Google\u2019s generative AI consumer products, and are imperceptible to humans \u2013 but can be detected by SynthID\u2019s technology,\u201d says a page on the site for DeepMind, Google\u2019s AI division.<\/p>\n Google presents SynthID as having what in the realm of digital watermarking is known as \u201crobustness\u201d \u2013 it claims to be able to detect the watermarks even if an image undergoes modifications, such as cropping or compression. Therefore, an image manipulated with Google\u2019s AI should contain detectable watermarks even if it has been saved multiple times or posted on social media.<\/p>\n <\/p>\n Google steers<\/a> those who want to use SynthID toward its Gemini AI chatbot, which they can prompt with questions about the authenticity of digital content.<\/p>\n \u201cWant to check if an image or video was generated, or edited, by Google AI? Ask Gemini,\u201d the SynthID landing page says.<\/p>\n We decided to do just that.<\/p>\n We saved the image file that the official White House account posted on X, bearing the filename G_R3H10WcAATYht.jfif<\/a>, and uploaded it to Gemini. We asked whether SynthID detected the image had been generated with Google\u2019s AI.<\/p>\n To test SynthID\u2019s claims of robustness, we also uploaded a further cropped and re-encoded image, which we named imgtest2.jpg.<\/p>\n Finally, we uploaded a copy of the photo where Levy Armstrong was not crying, as previously posted by Noem. (In the above screenshot, Gemini refers to Noem\u2019s photo as signal-2026-01-22-122805_002.jpeg because we downloaded it from the Signal messaging app).<\/p>\n \u201cI\u2019ve analyzed the images you provided,\u201d wrote Gemini. \u201cBased on the results from SynthID, all or part of the first two images were likely generated or modified with Google AI.\u201d<\/p>\n \u201cTechnical markers within the files imgtest2.jpg and G_R3H10WcAATYht.jfif indicate the use of Google\u2019s generative AI tools to alter the subject\u2019s appearance,\u201d the bot wrote. It also identified the version of the image posted by Noem as appearing to \u201cbe the original photograph.\u201d<\/p>\n <\/p>\n \n By signing up, I agree to receive emails from The Intercept and to the Privacy Policy<\/a> and Terms of Use<\/a>.<\/p>\n<\/p><\/div>\n<\/p><\/div>\n<\/p><\/div>\n<\/p><\/div>\n<\/div>\n <\/p>\n With confirmation from Google that its SynthID system had detected hidden forensic watermarks in the image, we reported in our story that the White House had posted an image that had been doctored with Google\u2019s AI.<\/p>\n This wasn\u2019t the only evidence the Whie House image wasn\u2019t real; Levy Armstrong\u2019s attorney told us that he was at the scene during the arrest and that she was not at all crying. The White House also openly described the image as a meme.<\/p>\n A few hours after our story published, Google told us that they \u201cdon\u2019t think we have an official comment to add.\u201d A few minutes after that, a spokesperson for the company got back to us and said they could not replicate the result we got. They asked us for the exact files we uploaded. We provided them.<\/p>\n The Google spokesperson then asked, \u201cWere you able to replicate it again just now?\u201d<\/p>\n We ran the analysis again, asking Gemini to see if SynthID detected the image had been manipulated with AI. This time, Gemini failed to reference SynthID at all \u2014 despite the fact we followed Google\u2019s instructions and explicitly asked the chatbot to use the detection tool by name. Gemini now claimed that the White House image was instead \u201can authentic photograph.\u201d<\/p>\n It was a striking reversal considering Gemini previously said that the image contained technical markers indicating the use of Google\u2019s generative AI. Gemini also said, \u201cThis version shows her looking stoic as she is being escorted by a federal agent\u201d \u2014 despite our question addressing the version of the image depicting Levy Armstrong in tears.<\/p>\n Less than an hour later, we ran the analysis one more time, prompting Gemini to yet again use SynthID to check whether the image had been manipulated with Google\u2019s AI. Unlike the second attempt, Gemini invoked SynthID as instructed. This time, however, it said, \u201cBased on an analysis using SynthID, this image was not made with Google AI, though the tool cannot determine if other AI products were used.\u201d<\/p>\n Google did not answer repeated questions about this discrepancy. In response to inquiries, the spokesperson continued to ask us to share the specific phrasing of the prompt that resulted in Gemini recognizing a SynthID marker in the White House image. <\/p>\n We didn\u2019t store that language, but told Google it was a straightforward prompt asking Gemini to check whether SynthID detected the image as being generated with Google\u2019s AI. We provided Google with information about our prompt and the files we used so the company could check its records of our queries in its Gemini and SynthID logs.<\/p>\n \u201cWe\u2019re trying to understand the discrepancy,\u201d said Katelin Jabbari, a manager of corporate communications at Google. Jabbari repeatedly asked if we could replicate the initial results, as \u201cnone of us here have been able to.\u201d<\/p>\n After further back and forth following subsequent inquiries, Jabbari said, \u201cSorry, don\u2019t have anything for you.\u201d<\/p>\n Aside from Google\u2019s proprietary tool, there is no easy way for users to test whether an image contains a SynthID watermark. That makes it difficult in this case to determine whether Google\u2019s system initially detected the presence of a SynthID watermark in an image without one, or if subsequent tests missed a SynthID watermark in an image that actually contains one.<\/p>\n As AI become increasingly pervasive, the industry is trying to put behind its long history of being what researchers call a \u201cbullshit generator<\/a>.\u201d<\/p>\n Supporters of the technology argue tools that can detect if something is AI will play a critical role establishing the common truth amid the pending flood of media generated or manipulated by AI. They point to their successes, as with one recent example where SynthID debunked an arrest photo of Venezuelan President Nicolas Maduro flanked by federal agents as an AI-generated<\/a> image. The Google tool said the photo was bullshit.<\/p>\n If AI-detection technology fails to produce consistent responses, though, there\u2019s reason to wonder who will call bullshit on the bullshit detector.<\/p>\n<\/p><\/div>\n #Googles #Detection #FlipFlops #Doctored #White #House #Photo<\/p>\n","protected":false},"excerpt":{"rendered":" When the official White House … <\/p>\n","protected":false},"author":1,"featured_media":15407,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[246],"tags":[],"_links":{"self":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/posts\/15406"}],"collection":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=15406"}],"version-history":[{"count":0,"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/posts\/15406\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/media\/15407"}],"wp:attachment":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=15406"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=15406"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=15406"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}![\"\"](\"https:\/\/theintercept.com\/wp-content\/uploads\/2026\/01\/Picture1.jpg?fit=936%2C452\")
Initial SynthID Results<\/h2>\n
\n We\u2019re independent of corporate interests \u2014 and powered by members. Join us. <\/h2>\n
\n Become a member
\n <\/a>\n <\/div>\n\n
\n Join Our Newsletter <\/span>
\n
\n Thank You For Joining! <\/span>
\n <\/h2>\n
\n Original reporting. Fearless journalism. Delivered to you. <\/span>
\n
\n Will you take the next step to support our independent journalism by becoming a member of The Intercept? <\/span>\n <\/p>\n
\n Become a member
\n <\/a><\/p>\nA Striking Reversal<\/h2>\n
![\"\"](\"https:\/\/theintercept.com\/wp-content\/uploads\/2026\/01\/Picture2.jpg?fit=936%2C290\")
![\"\"](\"https:\/\/theintercept.com\/wp-content\/uploads\/2026\/01\/Picture3.jpg?fit=936%2C656\")
Bullshit Detector?<\/h2>\n