Scores by kind of claim
Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.
-
Factual claims
Statements about the present or the past, checked against the record: did they hold up?
72 65 79 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
66 54 78 -
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
Insufficient record: 1 promise resolved, 5 needed for a score.
All types combined: 71 (likely 64 to 77, high confidence), from 79 claims: 60 factual claims, 19 predictions. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Balaji Srinivasan's claims
251 claims · page 3 of 3
“This appears to have been created by good old video editing tools rather than neural networks, but seems fake regardless.”
Balaji Srinivasan, writing on X 22 December 2019 Post by @balajis on X, 2019-12-22 (X)
A video circulating in December 2019 appears to have been made with conventional video editing tools rather than neural networks, and appears fake.
Unresolvable The public record could not settle it.
What would count as true or false
Judged as of 2019-12-22. The claim refers to the specific video Srinivasan was discussing in December 2019. True if forensic analysis, fact-checks or the creator's own account show the video was fake or manipulated, and that it was made with conventional editing (cuts, splicing, speed changes, compositing) rather than AI or deepfake methods. Mostly true if it was fake and mostly conventionally edited, with minor AI-assisted elements. Mixed if it was fake but made largely with neural or deepfake methods, or if it was conventionally edited but only misleadingly framed rather than fake. Mostly false if it was only mildly edited and largely authentic. False if it was authentic and unedited. The claim is hedged ("appears"), so a reasonable inference consistent with the later evidence is enough for true.
Evidence
- Welcome to the age of uncertainty POLITICO Europe · 17 December 2019
“The researchers found no evidence the clip of Bongo’s address was doctored, but they also couldn’t definitely rule out the possibility that it was.”
- Welcome to the age of uncertainty POLITICO Europe · 17 December 2019
“Both videos were examples of so-called cheap fakes, created with basic video-editing tools. Quickly debunked, they nonetheless dominated the global news cycle for several days.”
- Welcome to the age of uncertainty POLITICO Europe · 17 December 2019
“In May, a video of Speaker of the U.S. House of Representatives Nancy Pelosi, one of President Donald Trump’s most outspoken opponents, went viral. It had been slowed down to make the 79-year-old look drunk.”
Reasoning
None of the evidence identifies the video Srinivasan was discussing on 2019-12-22. E2 and E3 describe other viral videos from earlier in 2019, such as the slowed-down Pelosi clip, as cheap fakes made with basic editing tools. E1 covers the Bongo address, which researchers could neither confirm nor rule out as doctored. None of these can be tied to the video in the claim, so the evidence cannot show whether that video was fake or how it was made.
Cite this claim
Balaji Srinivasan, “A video circulating in December 2019 appears to have been made with conventional video editing tools rather than neural networks, and appears fake.” (said 22 December 2019). Verdict: unresolvable. Reliability API, claim c_f45fe2e45c52, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_f45fe2e45c52
Claims without a verdict
- Unresolvable 148
- Not yet due 21
- Own records only 2
- Not researched 227
- Not checkable 1,091
Is a verdict here wrong, or do you speak for Balaji Srinivasan? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.