By May 2023, ChatGPT had over 1.86B visits worldwide, USA ranking first in traffic contributors, followed by India. Every day, we come across numerous announcements from individuals across various fields, enthusiastically sharing how they have integrated AI tools into their respective creations. The latest trend at all sorts of receptions and conferences is to begin your presentation with “Esteemed guests, my following speech was generated by ChatGPT.”´The usage of AI tools has become so common that people do not make much out of it anymore. On the other hand, the ominous opinions sound louder than ever, warning people to approach with the utmost caution and suggest implementing artificial intelligence capabilities in governance, business, or science most conservatively.
Full of digital might
Estonia has stated in its Digital Agenda 2030 the importance of artificial intelligence as an accelerator of our digital governance – ‘full of digital might’ thanks to data-based solutions. Today, Estonia has implemented over 80 AI projects in 40 different public sector institutions to enhance the user experience in public services, raise the accessibility to them and turn governance more effective. According to the Artificial Intelligence Strategy of Estonia 2022-2023, over 20M euros will be dedicated to broadening the usage of AI on a national scale.
If the objective is to rely more on emerging technology about which we might not yet know everything, it’s worthwhile to ask about people’s attitudes and readiness on this matter.
Do you trust AI?
A survey on Human attitudes towards artificial intelligence was conducted in Estonia in 2019. According to this, 49% of people replied they had heard or read about artificial intelligence to some extent, and 33% stated that they had not. When they were further asked whether AI should be implemented more or should we approach AI more cautiously, 67% were inclined towards the latter, stating that they preferred definitely or somewhat more careful usage of AI. If asked, ‘Would you trust something you do not fully fathom?’ I am confident you would be hesitant to put your complete confidence in it, too.
Prof Margit Sutrop, head of the Centre for Ethics at Tartu University, states that her research shows that when we talk about trusting AI, it means trusting the people who create, use, and govern AI. Thus, trust is not so much about technology as people and institutions.
Also, recent research from Södertörn University studied people’s attitudes towards automated decision-making in the public sector from Estonia, Sweden and Germany and stated that people’s attitudes towards specific uses differ significantly between the countries. The study also showed that citizens of Estonia and Sweden are more positive than those of Germany when it comes to automation. This could be due to cultural differences between the countries and variations in people’s level of trust in authorities.
What do people want?
As I began this article by showing how more people engage with tools based on artificial intelligence simply because it has become more approachable, hence good governance practice also means the inclusion of users and asking what they want. I am delighted to know Estonia’s government is explicitly asking its citizen’s opinions to find out what matters most to them.
While discussing critical digital services, what people value the most are:
- Usability: the service works at all times, 24/7
- User’s control: I can return to the previous step to make changes without having to restart the whole application process
- I receive the services and benefits I am entitled to without taking additional steps to get them
- I can choose what information I share with the service provider
- The service is mostly pre-filled for me, so my responsibility is to check the data and submit/confirm.
Estonia is known globally for its extensive digital services, with 99% of the public services available online. However, consuming government services digitally is always voluntary, never an obligation. In the choice of the preferred channel, we see the main divide based on the complexity of the problem. Users with insufficient information or are dealing with a complex question would much instead contact a civil servant (either e-mail, call or go to the office). Users who solve a relatively easy task and are informed about it prefer self-service and chatbots.
Bürokratt is the interoperable network of chatbots implemented on the web pages of Estonia’s public institutions, a virtual assistant guiding a person directly to the relevant information or service. Tracing back to the question of trust, citizens state that they are aware of risks in technology related to broader usage of AI; 68% of them ‘do not see a significant obstacle or threat to use a virtual assistant like Bürokratt’.
However, they note that for them to turn to that channel, most likely, it has to offer an understandable value, like
- If Bürokratt will substantially decrease the time I spend on consuming government services compared to other channels
- If Bürokratt functions on all my devices
- I get a comprehensive overview of all the services or benefits I am entitled to and have an obligation to
- I am provided with an understandable description of how my data security and privacy are protected in Bürokratt.
That shortly sums up people’s motivation and will. Emerging technologies are a threat and a possibility at the same time. As long as it significantly improves the users’ life or experience, they are prone to try it. As legislators, civil servants and service providers, we must want to offer AI-backed tools ethically, not betray the people’s trust.