Headquarters: Munich, Germany
Assessed: AI behavioural analysis from video, positioned for coaching and training.
Rubric: version 1. Search protocol applied and all sources accessed: 2026-09-22.
Documented 0, partial 4, absent 4, of eight criteria. This records what is published about the validity and fairness of the scores. It is not an assessment of the product, the model or the underlying science.
| # | Criterion | Level |
|---|---|---|
| C1 | Technical manual or validity report publicly available without a sales conversation | P |
| C2 | Validity coefficients reported with sample size and uncertainty | A |
| C3 | Adverse impact reported by named subgroup | P |
| C4 | Scored construct named and defined, not described only by outcome | P |
| C5 | Evidence of comparability across languages, versions or delivery modes | A |
| C6 | Human oversight documented, with a candidate route to contest a score | P |
| C7 | Independent review named and the reviewer identified | A |
| C8 | Reliability reported, and reported by group | A |
C1. Technical manual or validity report publicly available without a sales conversation
Partial.
"To dive deeper, read our whitepaper – We believe in science and transparency", with "Download whitepaper" and "Download now" calls to action routing to an email-capture form.
Source: https://www.retorio.com/en/scienceai (accessed 2026-09-22).
The document requires contact details and was not read.
C2. Validity coefficients reported with sample size and uncertainty
Absent.
No public source was found for this criterion by the search protocol.
Pages examined include https://www.retorio.com/en/scienceai (accessed 2026-09-22).
The figures reported are detection accuracy for facial and verbal features, not criterion-related validity.
C3. Adverse impact reported by named subgroup
Partial.
"Our AI has been debiased and tested on global datasets, incl. the UCLA Fairface dataset"
Source: https://www.retorio.com/en/ethical-data (accessed 2026-09-22).
A named external dataset is used for bias testing. No selection rate, impact ratio or subgroup result is published.
C4. Scored construct named and defined, not described only by outcome
Partial.
"Behavioral Intelligence", described as analysing "facial expressions, gestures, tone of voice, and even language".
Source: https://www.retorio.com/en/scienceai (accessed 2026-09-22).
A proprietary construct name, with a description of the inputs rather than a definition.
C5. Evidence of comparability across languages, versions or delivery modes
Absent.
No public source was found for this criterion by the search protocol.
Pages examined include https://www.retorio.com/en/scienceai (accessed 2026-09-22).
C6. Human oversight documented, with a candidate route to contest a score
Partial.
"Automated individual decisions about whether or not an application should be considered are not possible … the human still has the authority to make decisions."
Source: https://www.retorio.com/en/ethical-data (accessed 2026-09-22).
No route to query or contest a score is described.
C7. Independent review named and the reviewer identified
Absent.
No public source was found for this criterion by the search protocol.
Pages examined include https://www.retorio.com/en/ethical-data (accessed 2026-09-22).
The reference is to data sources rather than to a review of the system.
C8. Reliability reported, and reported by group
Absent.
No public source was found for this criterion by the search protocol.
Pages examined include https://www.retorio.com/en/scienceai (accessed 2026-09-22).
Absent records the outcome of the published search protocol on 2026-09-22. Where a source is supplied that the search did not reach, the cell is re-scored under the corrections policy, with the date of the change recorded and the previous level retained.
Back to the index, the rubric and the methods note