Post by @SponceyM
81% credible (86% factual, 71% presentation). Herasight's CogPGT IQ predictor claims align with polygenic research, supported by validation from large cohorts like UK Biobank. However, the presentation omits critical ethical issues and limitations in predictive accuracy across ancestries, resulting in a promotional overstatement of benefits.
Analysis Summary
Herasight has unveiled CogPGT, claiming it as the world's most powerful genetic predictor of IQ with a population correlation of 0.51 and within-family correlation of 0.45, enabling up to a 9-point IQ boost via embryo selection in IVF. The claims align with recent advancements in polygenic scoring, though real-world efficacy remains debated due to environmental factors and ethical issues. Validation draws from large cohorts like UK Biobank, but counterarguments highlight limitations in predictive accuracy across ancestries and potential societal risks.
Original Content
The Facts
The core scientific claims about CogPGT's correlations are supported by the company's validation whitepaper and align with emerging polygenic research, though the 'world's most powerful' assertion is promotional and correlations may not fully translate to causal IQ gains due to gene-environment interactions. Mostly Accurate, with caveats on overstatement of benefits and omission of risks like reduced genetic diversity or ethical concerns in embryo selection.
Benefit of the Doubt
The post advances a pro-innovation agenda for Herasight's polygenic embryo screening technology, emphasizing scientific breakthroughs and practical benefits to attract IVF customers and normalize IQ selection. It highlights positive correlations and IQ boosts while omitting critical counterarguments such as ethical dilemmas (e.g., eugenics concerns), risks of inaccurate predictions across diverse ancestries, and potential societal inequalities from access disparities. This selective framing shapes perception toward optimism, downplaying controversies to build excitement around the tool.
Predictions Made
Claims about future events that can be verified later
Herasight customers can boost the expected IQ of their children by up to 9 points by selecting the embryo with the highest CogPGT score
Prior: 50% (polygenic selection predictions vary, with gene-environment interactions reducing efficacy). Evidence: Author track record factual, expertise relevant, verified by X posts and Substack; image shows ~10 IQ spread for Europeans; caveats on ancestry biases. Posterior: 75%.
Visual Content Analysis
Images included in the original content
VISUAL DESCRIPTION
The image features two charts: the left is a bar graph comparing population (light blue bar at 0.51) and within-family (darker blue bar at 0.45) correlations of CogPGT with general cognitive ability (g); the right is a line graph showing expected IQ spread for European (solid black line with dots), East Asian (dashed gray line), and African (solid orange line with dots) ancestries across 1-20 embryos, with lines rising from near 0 to around 10-15 IQ points. A Herasight logo appears at the bottom.
TEXT IN IMAGE
CogPGT 1.0 correlation with general cognitive ability (g) r 0.5 0.51 (Population, blue bar) 0.4 0.3 0.2 0.1 0.45 (Within-family, darker blue bar) Population Within-family Expected IQ spread predicted by CogPGT 1.0 IQ spread 20 15 10 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Number of embryos Lines: European (black), East Asian (gray), African (orange dots) Validated within-family on 6,442 siblings in the UKB and 736 siblings in the ABCD cohort. Herasight
MANIPULATION
No signs of editing, inconsistencies, or artifacts; charts appear professionally generated with consistent styling, axes, and data points typical of scientific visualizations.
TEMPORAL ACCURACY
The image references recent validation data from UK Biobank (UKB) and ABCD cohort, aligning with the 2025 announcement date; no outdated elements like old cohort references.
LOCATION ACCURACY
The image is a data visualization without specific locations; no geographical claims are made, so spatial framing is not applicable.
FACT-CHECK
The charts match descriptions in Herasight's announcement and whitepaper; correlations and IQ spread predictions are consistent with polygenic scoring literature, though lower performance for non-European ancestries reflects known GWAS biases. No contradictions found in reverse image or web searches.
How Is This Framed?
Biases, omissions, and misleading presentation techniques detected
Selective presentation omits ethical dilemmas, prediction limitations across ancestries, and societal risks, altering interpretation toward unmitigated benefits.
Problematic phrases:
"We achieve a correlation with IQ of 0.51""boost the expected IQ of their children by up to 9 points"What's actually there:
Correlations supported but do not guarantee causal gains due to environmental factors
What's implied:
Straightforward and reliable IQ enhancement via selection
Impact: Readers undervalue risks and controversies, perceiving the technology as a simple positive advancement.
Implies direct causal link between correlation scores and IQ increase without substantiating causation over correlation.
Problematic phrases:
"by selecting the embryo with the highest CogPGT score"What's actually there:
Population correlation of 0.51, within-family 0.45
What's implied:
Selection directly causes up to 9-point IQ boost
Impact: Misleads readers into believing predictive scores translate unproblematically to real-world cognitive gains.
Uses 'today' to create immediacy for a product announcement, heightening perceived novelty and excitement.
Problematic phrases:
"Today we reveal"What's actually there:
Standard product launch
What's implied:
Breaking, urgent scientific breakthrough
Impact: Increases reader excitement and perceived timeliness, encouraging quick positive judgment without reflection.
Highlights maximum 'up to 9 points' boost and 'world's most powerful' without average outcomes or full comparative data, exaggerating magnitude.
Problematic phrases:
"up to 9 points""the world’s most powerful"What's actually there:
Emerging research shows variable gains, not universally highest
What's implied:
Guaranteed superior performance and substantial average boost
Impact: Inflates perceived effectiveness and superiority, downplaying variability and need for broader context.
Sources & References
External sources consulted for this analysis
https://herasight.substack.com/p/cogpgt
https://www.nejm.org/doi/full/10.1056/NEJMsr2105065
https://academic.oup.com/humupd/article/30/5/529/7684172
https://pmc.ncbi.nlm.nih.gov/articles/PMC6957074/
https://hms.harvard.edu/news/study-reveals-public-opinion-polygenic-embryo-screening-ivf
https://www.sciencedirect.com/science/article/abs/pii/S0277953624000431
https://www.theguardian.com/science/2021/oct/17/polygenic-screening-of-embryos-is-here-but-is-it-ethical
https://herasight.substack.com/p/cogpgt
https://nepc.colorado.edu/blog/educational-genomics
https://www.vox.com/future-perfect/459003/designer-babies-embryo-selection-polygenic-testing-ethics
https://www.washingtonpost.com/outlook/2021/06/14/genetic-screening-ivf-moral-dilemmas/
https://www.mdpi.com/2073-4425/12/8/1105/htm
https://www.statnews.com/2021/07/01/researchers-raise-concerns-genetic-risk-scores-pick-healthier-embryos/
https://www.sciencedirect.com/science/article/abs/pii/S0277953624000431
https://x.com/SponceyM/status/1980660198441447568
https://x.com/SponceyM/status/1980660204162412876
https://x.com/SponceyM/status/1980766174083461198
https://x.com/SponceyM/status/1980660207098491073
https://x.com/SponceyM/status/1980660201536782422
https://x.com/SponceyM/status/1980660224307716486
https://herasight.substack.com/p/cogpgt
https://en.wikipedia.org/wiki/Heritability_of_IQ
https://pmc.ncbi.nlm.nih.gov/articles/PMC5985927/
https://pmc.ncbi.nlm.nih.gov/articles/PMC12416016/
https://pmc.ncbi.nlm.nih.gov/articles/PMC6986352/
https://consensus.app/home/blog/is-our-maximum-iq-determined-by-our-genetics/
https://www.theguardian.com/science/2024/oct/18/what-is-genomic-prediction-and-can-embryos-really-be-screened-for-iq
https://di.aerzteblatt.de/int/archive/article/242633
https://jkms.org/DOIx.php?id=10.3346%2Fjkms.2025.40.e132
https://www.benzinga.com/news/topics/25/08/47034820/can-ai-predict-your-babys-intelligence-new-genetics-startup-herasight-says-yes-sparking-fears-of-eugenics-2-0-in-ivf
https://consensus.app/home/blog/are-iq-scores-a-good-predictor-of-general-intelligence/
https://www.psypost.org/groundbreaking-study-reveals-the-impact-of-genetics-on-iq-scores-over-time/
https://link.springer.com/article/10.1007/s00392-020-01721-x
https://nature.com/articles/s41562-020-00952-2
https://x.com/SponceyM/status/1980660198441447568
https://x.com/SponceyM/status/1980660204162412876
https://x.com/SponceyM/status/1980766174083461198
https://x.com/SponceyM/status/1980660207098491073
https://x.com/SponceyM/status/1980660224307716486
https://x.com/SponceyM/status/1980660201536782422
Want to see @SponceyM's track record?
View their credibility score and all analyzed statements