New paper out. 'Tracking historical changes in trustworthiness using machine learning analyses of facial cues in paintings' @naturecom with @CoralieChevall1 @JulieGrezes and Lou Safra
https://tinyurl.com/yxzqzb4r 

Building on recent advances in social cognition, we design an algorithm to automatically generate trustworthiness evaluations for the facial action units (smile, eye brows, etc.). https://tinyurl.com/y4jvat74 

The algorithm generates automatic human-like trustworthiness ratings on portraits based on the muscle contractions (facial action units) detected in facial displays using the open software #OpenFace

We used it on historical portraits. One can display all kind of qualities in a portrait: dignity, strenght, wealth, youth, and also trustworthiness. Quantifying trustworthiness tells us how much looking nice and trustworthy is important in a society.

For instance, Descartes' famous portrait could have been very different.

To assess the generalizability of our model, we then tested its validity on four databases of natural faces rated by real participants. We first demonstrated that the algorithm produced ratings that were aligned with those produced by human participants in 4 databases.

We then checked that the algorithm was susceptible to the same biases as humans, i.e., rating younger, feminine, and happy faces as more trustworthy.

We replicated all these findings outside well-controlled databases by analyzing all the images (photographs and paintings) obtained from a Google image search for ‘women portraits’ vs ‘male portraits’.

To assess the evolution of trustworthiness displays in history, we first analyzed the paintings of the @smithsoniannpg the largest online database of historical portraits (analyzed N = 1962 English portraits from 1505 to 2016).

We found a significant increase of trustworthiness displays with time, which makes sense if you think about it (we did not use photographies, but we just run this comparison for fun: Here are the two Queen Elisabeth, in two different periods)

Since we are having fun... Another comparison, controlling for age (Elisabeth I: -0.53 ; Megan Markle: 1.88)

We then replicated our findings on the Web Gallery of Art an important fine art repository (N = 4106 portraits) spanning 19 Western European countries seven centuries (1360–1918).

Again, we found a significant increase in trustworthiness displays with time.

Whether such increased in trustworthiness in portraits parallels an actual shift in social trust remains an open question. To assess the validity of this assumption, we applied our algorithm to selfies posted on Instagram in six cities around the world. Thanks @selfiecity

We found that people located in places where interpersonal trust and cooperation are higher (as assessed in the European and World Value Surveys) displayed higher levels of trustworthiness in their selfies.

Our results thus show that trustworthiness in portraits increased over the period 1500–2000 paralleling the decline of interpersonal violence and the rise of democratic values observed in Western Europe. @DeirdreMcClosk @sapinker

Another open question is that of the potential predictors of trustworthiness fluctuations in social displays. We first examined the role of resources.

Because losses have more dramatic effects for poorer individuals, individuals with lower resources are arguably more exposed by exploitation risks and should therefore have lower levels of social trust.

In line with this reasoning, international surveys show a strong association between resources and social trust. @OurWorldInData

Our analysis of the National Portraits Gallery database revealed an association between higher levels of affluence and higher levels of trustworthiness displays between the 16th and the 21st centuries. Same for Web Gallery of Art.

Demonstrating that the association between GDP and the rise of trustworthiness is causal would of course require additional data. However, we were able to investigate the dynamics of these historical changes by running time-lag analyses on trustworthiness and GDP per capita.

We found that changes in GDP per capita predicted future changes in trustworthiness displays in our two databases two decades later. Importantly, changes in trustworthiness displays did not predict future changes in GDP per capita.

All together, these findings complement existing qualitative historical accounts and demonstrate how insights from cognitive sciences can enrich our understanding of history And since you read all this thread, you surely deserve one last comparison! (Biden : 1.39; Trump : -0.63)