Estonian residents support using artificial intelligence in state services

Tea Danilov, Head of the Foresight Centre, said that the Estonians’ favourable stance is primarily based on the general trust in technology and the digital state. “Estonian residents trust using artificial intelligence in state services more than residents of Sweden and Germany, but their actual awareness of the operating mechanisms and specificities of automated decision-making processes is low,” said Danilov. “I would say this is a kind of initial support and lasting support still needs to be earned. As various studies suggest, to ensure the legitimacy and trustworthiness of automatic decisions, citizens must be able to understand and influence the decisions, for example, by explaining or challenging their cases.”

Automated Decision-Making (ADM) means that tasks are completed and decisions are made based on algorithms that use specific rules or models with learning capabilities, such as predictive data analysis.

The survey conducted in three European countries found that Estonian residents are in favour of ADM-made decisions and every second Estonian resident expects savings in cost and time from it. This group is characterised by high trust in institutions and low risk awareness. Nearly 30% of the population, however, are classified as a group with high awareness of data owner rights and data protection and low trust in state institutions. Around a third of the respondents could not or did not want to choose.

Previous studies have shown that transparency and broad citizen participation are critical in ensuring trust in ADM. There is a significant fear that needs to be addressed that automated decision-making systems have limited capability to take individual specifics into account or make exceptions that could be important in making fair decisions. “For example, ADM is not able to consider the wider context and make exceptions as artificial intelligence lacks emotions, such as sense of justice or empathy, which are characteristic of humans,” said Danilov.

In the case of decisions made by artificial intelligence, Estonians who participated in the comparative study of the countries value efficiency, protection of personal interests and solidarity higher than Swedish and German respondents. Estonian respondents rated the suitability of ADM in public administration higher (2.5 in Estonia vs 2.2 in Sweden and Germany). Estonians’ optimism in this regard is notable, considering their low awareness of ADM systems (score 1.6 on a scale of 1-5).

The short report  (in Estonian) ‘Estonian residents’ readiness for automated decision-making in public sector’ is part of the Foresight Centre’s research stream ‘Future of the personal state’. The research stream aims to explain the possibilities of data-based personalisation in social and education support and services and create alternative scenarios.