Fact check
“About 70% of the training data for the encoder/decoder comes from images not paired with fMRI scans.”
Reasoning
The provided excerpt states that "around 70% of the training data is from images that were not originally paired with fMRI scans," which matches the claim. No contradictory evidence is presented.
On confidence: The claim is directly supported by a reputable source (MIT Technology Review) quoting a researcher who states the approximate proportion.
Important context
The figure is an approximate value ('around 70%') and refers to the training data used for the encoder/decoder models in the described AI mind‑reading system.
Evidence
Supporting (1)
- Tier 1 — Primary sourceAn AI “mind-reading” tool can reconstruct what you’re looking at based on a brain scan | MIT Technology Review
to train their models on as many images as they want, even though they might never have been shown to a person in an fMRI scanner. Irani says that around 70% of the training data is from images that were not originally paired with fMRI…
Limitations
The evidence is a single secondary source summarizing a researcher’s statement; the original study or dataset details are not provided for independent verification. No independent origin corroborated this claim, so it cannot be verified and confidence cannot be high.
- Last verified:
- Oct 1, 2026, 6:38 AM CDT
- Pipeline:
- 2.1.0
- Claim type:
- Factual
Where this claim appeared
An AI “mind-reading” tool can reconstruct what you’re looking at based on a brain scanMIT Technology Review · Jessica Hamzelou