AI, trust, deepfakes and epistemic resilience. 'Prepare, Don't Panic' initiative on media manipulation/generative AI. TED speaker deepfakes. Former ED WITNESS.
New York, NY
Joined February 2008
- Tweets32.9K
- Following11.8K
- Followers11.7K
- Likes34.1K
Sam Gregory retweeted
One of the red flags of AI-written text is the heavy use of these odd linguistic constructions called “cataphoric teasers” to create an artificial feeling of suspense. These are easy to spot because they usually take the form of phrases like, “Here’s the part that nobody tells you…” or “Here’s what most people get wrong…” or “Here’s where it gets interesting.” From my experience, text written by Claude is often saturated with these types of phrases. I don't see it as much in ChatGPT (that's anecdotal, though. I might be wrong).
I also thought @paulbloomatyale’s reflections about this were good: AI writing often has a clickbait-like urgency, as if every sentence was trying to keep you reading substack.com/@smallpotatoes/…
Sam Gregory retweeted
"We take misinformation seriously... [so we decided destroy trust in a key reference point in the epistemology of journalists, fact checkers and researchers]"
Well done, you put Synth ID in it. Now fact-checkers can write an article after it's used to create a viral fake story.
We take misinformation seriously – every image created with Nano Banana in Google Earth includes the SynthID digital watermark, so if someone is unsure about an image, they can ask the Gemini app or use Lens in Search to see if the image was AI-generated. In addition, we prevent image creation on harmful topics and are continually updating our protections → policies.google.com/terms/ge…
Sam Gregory retweeted
How can journalists navigate the rising threat of deepfakes and synthetic media?
@SamGregory introduces the Deepfakes Rapid Response Force, a project from @witnessorg that connects reporters with forensic experts to verify suspicious content and protect reporting integrity.
Watch the full Lightning Talk from #MediaParty 2025 now available here👇
mediaparty.org/2026/04/16/sa…
Sam Gregory retweeted
Iran is the worst-case scenario for everything we’ve been warning about: AI fakes, civilian footage dismissed as fabricated, and underneath it all, a regime that has engineered the doubt itself.
“The problem is curated access. The regime controls who gets to document and what gets documented.”
I spoke to @morganmeaker for Bloomberg Weekend on the information war in the Iran war. bloomberg.com/news/articles/…
Sam Gregory retweeted
New piece with shirin anlen based on our work at @witnessorg on how 'forensic cosplay' (h/t @CraigSilverman) is being used to discredit authentic photos of civilian casualties in the Iran war.
The fake heatmap and "ELA" case spread through my own networks. People I know concluded the photos were fake, primed by justified distrust of state sources. That distrust is legitimate. But it can lead to dangerous denialism.
With references to @Shayan86 and @talhagin who have been working on the deluge of AI cases during this war and insights from forensics specialists.
Sam Gregory retweeted
This past week I wrote about what we’re seeing now: evidence and documentation of what’s actually happening being actively undermined by a polluted info environment with AI accelerating the damage.
The spread of fake imagery of the Iran war is helping make the question “Is this real?” all but unanswerable, Mahsa Alimardani argues: theatlantic.com/ideas/2026/0…
Sam Gregory retweeted
Since the Iran-Israel war began in June, @WITNESSorg has been tracking how AI is reshaping the information environment around the conflict, starting with this piece with @samgregory and work that fed into a recent @OversightBoard ruling on governing AI content during war.
During the Iran-Israel war, civilians in Iran had little access to trustworthy information. AI deepfakes circulated online exacerbated this situation. What are the implications?
Mahsa Alimardani and @SamGregory explained: carnegieendowment.org/resear…
Sam Gregory retweeted
At the report launch, our Executive Director @SamGregory discussed the impact of AI on humanitarian action in an increasingly complex information environment and the need to maintain the ability to know what is real.
📲Read more here: blog.witness.org/2026/03/ifr…
Sam Gregory retweeted
In an AI-shaped world, video is still crucial for accountability, but its credibility is under pressure.
🔗Read the full @nytimes article where our Executive Director, @SamGregory, discusses this “deeply contradictory” moment and how we can still show what’s real with video: nytimes.com/2026/02/15/us/po…
Sam Gregory retweeted
I'm co-hosting a new BBC podcast! It's called The Interface, and it's all about how tech is rewiring your week and your world.
bbc.com/mediacentre/bbcstudi…
ALT Karen Hao
ALT Nicky Woolf
ALT Thomas Germain
Sam Gregory retweeted
Filming federal agents in public is legal, but avoiding a dangerous—even deadly—confrontation isn’t guaranteed. Here’s how to record ICE and CBP agents as safely as possible and have an impact. wired.com/story/how-to-film-…
Sam Gregory retweeted
Quilicura, Chile, one of the communities I wrote about in EMPIRE OF AI, has launched a brilliant initiative to inspire more responsible AI prompting. Today, don't use AI; ask the townspeople instead: quili.ai. So heartened to see this creative act of resistance.
Sam Gregory retweeted
WITNESS is hosting a public session in Arabic about AI-generated content.
📅 Feb 10 | ⏰ 1:00 PM Beirut time
تستضيف WITNESS جلسة تدريبية عامة باللغة العربية حول فهم واكتشاف
المحتوى المُنتَج بالذكاء الاصطناعي.
يمكن التقديم من خلال الرابط
📲wit.to/AI_WebinarArabic
Sam Gregory retweeted
.@JeffBezos I cover international security for @washingtonpost and reported on ISIS, Jamal Khashoggi’s murder, al-Qaeda, Iran and Israel. I’ve been imprisoned, lived under threats, and escaped kidnappings to do this work. #Journalism matters to America and the world. #SaveThePost
Sam Gregory retweeted
Russia banned me and colleagues for our reporting for @washingtonpost .This is the price of telling the truth.
.@JeffBezos, journalism is under fire. Please help all of #America and the #World by protecting #Journalism that is fair, independent, and truly matters.
#SaveThePost
Sam Gregory retweeted
"Separate analyses by The New York Times, The Wall Street Journal, The Associated Press, CBS News and other organizations all concluded that the videos contradict the Trump administration’s description of the killing."
(ht/ @nancylevine)
nytimes.com/2026/01/25/opini…
Sam Gregory retweeted
With BBC Verify colleagues, we made this video analysing bystander video footage of the moments before the killing of Alex Pretti by federal immigration officers in Minneapolis. bbc.co.uk/news/videos/cg4e0r…
