AI-Generated De Niro Image Posted on Trump’s Truth Social Account
A fake image of Robert De Niro in a pro-Trump shirt was posted on Trump's Truth Social. Detection tools rated it 99.5% AI-generated with an OpenAI SynthID watermark.
An image showing actor Robert De Niro wearing a "Trump Is My President" T-shirt was posted to the @realDonaldTrump account on Truth Social on September 6, 2026, with the caption "Even this jerk is beginning to get it!" The image is fake. Fact-checkers at Lead Stories ran it through two detection tools: Hive Moderation rated it 99.5% likely to be AI-generated, and the OpenAI Verify tool detected a SynthID watermark showing it was made with OpenAI tools. There is no reporting that De Niro has changed his position on Trump.
What happened
| Detail | Fact |
|---|---|
| Account | @realDonaldTrump on Truth Social |
| Date posted | September 6, 2026 |
| Caption | “Even this jerk is beginning to get it!” |
| Hive Moderation rating | 99.5% likely AI-generated |
| Likely generation model (Hive) | GPT-4o (78.5% probability) |
| OpenAI Verify result | SynthID watermark detected, confirming OpenAI tools |
The post showed Robert De Niro apparently wearing a shirt with the words “Trump Is My President.” On its face, it implied the actor, a longtime vocal critic of Donald Trump, had changed his stance. He has not, according to searches of Yahoo News and Google News, which returned no reporting of any such shift.
Lead Stories, the fact-checking outlet that investigated the image, used two separate tools. Hive Moderation is a content detection service that scores images on the likelihood they were machine-generated rather than photographed. OpenAI Verify is a tool that reads SynthID watermarks, invisible digital signatures that OpenAI embeds in images created with its models. Both pointed to the same conclusion.
Why does this matter for businesses and marketers?
This story is a clean demonstration of how easy it now is to fabricate a convincing celebrity endorsement. A realistic AI image, a short caption, and a large social media account is all it takes to generate news coverage and viral shares before a correction catches up.
For any business running social media advertising or managing a brand’s public presence, the risk is real in two directions. Your brand could be falsely associated with a figure or statement through a fabricated image. Or your competitors, or bad actors, could use similar tools to spread misleading content that affects your market.
Detection tools like Hive Moderation and OpenAI Verify are becoming more important as part of basic due diligence. If your team shares or responds to viral content, checking it against an AI detection tool before amplifying it should be a standard step. SynthID watermarks are also worth understanding: they are baked into images at generation time and survive most basic edits, making them a useful forensic signal even when an image looks convincing.
The broader issue is that AI image generation, including the GPT-4o model flagged here, is now accessible to anyone. The same tools businesses use to create social media content can be used to fabricate it. Knowing how these tools leave traces is genuinely useful knowledge.
Our take
The detection result here is about as clean as it gets: 99.5% on Hive plus a confirmed SynthID watermark. That is not a close call. What is worth noting is that both tools reached the same conclusion independently, which is good practice. Relying on a single detector is not enough, since no tool is perfect and false positives exist.
What we find more interesting than the image itself is the watermark. OpenAI embeds SynthID signatures in images generated with its tools, meaning there is a verifiable trail back to the model. That is a meaningful accountability feature, but only if people know to look for it. Most don’t. As we’ve covered in OpenAI’s broader push toward disclosure frameworks, the infrastructure for accountability is being built, but adoption is slow.
If you are verifying images before sharing them, the workflow is straightforward: upload to Hive Moderation, then cross-check with OpenAI Verify. Two minutes of checking can save a significant amount of reputational risk.
What to do about it
- Before sharing any viral image involving a public figure, upload it to the Hive Moderation AI-Generated Content Detection Tool.
- Cross-check with the OpenAI Verify tool to look for a SynthID watermark.
- Search Google News and Yahoo News for the claim in the image before amplifying it on any channel.
- Brief your social media or content team on what SynthID watermarks are and how detection tools work, even a short internal note reduces risk.
One realistic image and one caption can move faster than any correction: build the checking habit before you need it.
Frequently asked questions
How can you tell if an image is AI-generated?
Tools like Hive Moderation score images on their likelihood of being machine-generated. OpenAI Verify checks for SynthID watermarks, which OpenAI embeds in images created with its tools. Using both together gives a more reliable result.
What is a SynthID watermark?
SynthID is an invisible digital watermark that OpenAI embeds in images generated with its models. It survives most basic edits and can be detected using the OpenAI Verify tool, revealing that the image was created with OpenAI software.
Was the Robert De Niro Trump shirt image real?
No. The image was rated 99.5% likely AI-generated by Hive Moderation, and the OpenAI Verify tool detected a SynthID watermark, confirming it was made with OpenAI tools. De Niro has not publicly changed his position on Trump.
Which AI model was used to make the De Niro fake image?
Hive Moderation flagged a 78.5% probability that the image was generated using GPT-4o, OpenAI's image generation model. The OpenAI Verify tool confirmed the presence of an OpenAI SynthID watermark.
