Nature Scientific Reports publishes our study on trust for the sharing ecomony

We present an evolutionary trust game, taking punishment and protection into consideration, to investigate the formation of trust in the so-called sharing economy from a population perspective. This sharing economy trust model comprises four types of players: a trustworthy provider, an untrustworthy provider, a trustworthy consumer, and an untrustworthy consumer. Punishment in the form of penalty for untrustworthy providers and protection in the form of insurance for consumers are mechanisms adopted to prevent untrustworthy behaviour.

Our results show that each player type influences the ‘existence’ and ‘survival’ of other types of players, and untrustworthy players do not necessarily dominate the population even when the temptation to defect (i.e., to be untrustworthy) is high. Additionally, we observe that imposing a heavier penalty or having insurance for all consumers (trustworthy and untrustworthy) can be counterproductive for promoting trustworthiness in the population and increasing the global net wealth. Our findings have important implications for understanding trust in the context of the sharing economy, and for clarifying the usefulness of protection policies within it.

Open access publication and PDF: https://www.nature.com/articles/s41598-019-55384-4

New publication on the use of AI for marketing (California Management Review)

Authors from US, the Netherlands, and Spain have published a practical guide for the adoption of AI technologies in Marketing, following the data science standard CRISP-DM. The paper was published in California Management Review (CMR) in July 2019, one of the most relevant publications in management (impact factor of 5.0).

CMR also published a blog entry about the publication and a brilliant promo video:

Main steps for developing and integrating an AI DSS for Word-of-mouth campaigns

In a nutshell, the authors point out that artificial intelligence works best when it is applied in careful steps:

  1. Define a clear business objective.
  2. Collect and sort an enormous amount of raw data.
  3. Choose the data for your project, and organize it into a digitally digestible format.
  4. Create a program that models how a human would use this data to reach the business objective defined in step one.
  5. Evaluate the outcome to see if the AI application has met the objective defined in step one.
  6. Deploy the AI application.

Massive media coverage about our study on the 11M terrorist attacks and elections’ influence

The results of our paper, recently published in the Knowledge-based Systems journal, have appeared in many Spanish newspapers:

La Vanguardia
El Diario.es
El Economista
El dia.es
IDEAL
Noticias de Gipuzkoa

Government, politicians, and mass media generated a large quantity of information after the bombing attacks in Madrid on the 11th of March 2004. This information had two competing dimensions on the terrorist group responsible for the attacks: ETA and Al’Qaeda. The framing theory could explain how this information influenced the Spanish national elections on the 14th of March, three days after the attacks. We propose to analyze this political scenario using agent-based modeling to recreate the environment and framing effect of the three days prior to the elections. Using our model we define several experiments where we observe how media communications influence agent voters after calibrating the model with real data. These experiments are what-if scenarios where we analyze alternatives for mass media communication messages and word-of-mouth behaviors. Our results suggest that the framing effect affected the election results by influencing voters. These results also outline the aggregated impact of mass media channels and the different role of each party segment of voters during this period.

See also press release of Canal UGR