Paperjam.lu

Xavier Bettel reiterated his willingness to take individual perceptions into consideration before moving forward with the national strategy related to artificial intelligence. 

Two years ago, during the presentation of the first national strategy on artificial intelligence, Bettel (DP), was more socialist than the socialist who accompanied him, Étienne Schneider (LSAP), placing people at the heart of developments.

Two years later, during the first public consultation on the same subject, the head of government is even less “liberal” from a dogmatic point of view, insisting on his government's desire to be the most “liberal” and inclusive possible. In just half an hour of a well-conducted press conference, Bettel only spoke the word "business" once. There is no question of freeing up the market, of nurturing opportunities for growth or improving productivity. The prime minister wants the population to adhere to developments in public policies on the subject.

Of the 19,993 residents identified and contacted by the State Information Technology Center for this public consultation conducted by the Luxembourg Institute of Socio-Economic Research (Liser), 2,383 responded. A good sample, in Luxembourg, varies between 500 and 1,000 people; 12% represents a good base.

The 10 lessons

So what did they say?

1. That 91% of them are connected to a high-speed network (fixed or mobile), that 9 out of 10 use their smartphone every day (versus 81% for a computer and 36% for connected, 13% for voice assistants and 12% for game consoles).

2. That they on social networks a lot (65%, compared to 48% for streaming sites and 31% for mapping applications).

3. That they are aware their data is being collected (94%) and that this "worries" them: more than half (54%) think that the collection of their personal data does not respect their privacy (one person in three cannot answer this question). 93% of those who feel that their privacy is not respected think that their data can be sold, 86% do not know why it is collected, 83% do not know what data is collected, 81% do not trust in those who collect the data and 74% believe their data is not being collected. They are also likely to think that their data is not protected enough and to know the European regulation on data protection. Less than one in three people think this is enough to protect them and 45% don't think so.

4. 43% of respondents monitor their lifestyle (food and sport) with an app, three-quarters to access public transport or avoid a traffic or roadworks, almost half admit that the tools for education are of no use to them, while a third (30%) have already benefited from adapted online training.

5. Almost 9 out of 10 residents use digital tools for public life, in particular to request official documents, and in two-thirds of cases on a regular basis, while few of them worry about environmental or energy issues: 61% do not seek to improve their energy consumption, two thirds do not use GouvAlert.lu and 79% do not use Meng Loft.

6. Among those who have a job, a third say that Linkedin or Viadeo got them a better job and the same proportion said that the sites were of no use to them. One in four was able to find new clients, patients, users.

7. 70% believe artificial intelligence facilitates the tasks of daily life and 64% that it saves them from repetitive tasks in the professional world. 22% say they don't have enough knowledge to have an opinion on the matter.

8. 70% of respondents say AI can act in a discriminatory or biased manner, 64% that it is unreliable and 56% that it is unable to distinguish between positive and negative consequences; more than one in two does not know what it is used for and with what data.

9. 58% of people trust AI when it is implemented by the government, compared to 41% by the private sector.

10. 84% of respondents believe AI will be useful in the security and defence sector, a figure that then drops to 34% for mobility and 13% for the environment. The risk outweighs the benefits for them in health, education and industry. The gap is largest in health, with 30% seeing benefits and 51% seeing risks. But they think it should be used in disease prevention (69%), for medical diagnosis (72%) or to obtain better adapted treatments (67%).

Structure to keep AI on track 

The prime minister also lingered for a minute on the need to set up an ethics committee to oversee future applications of AI within the state. Mr. Bettel indicated that discussions were taking place, internally, among officials who are confronted daily with these issues and that the state council would be consulted before the creation of a new body or a new division at the national ethics commission that exists.

A subject that could evolve, as this issue is also mentioned in the European Commission's draft regulation on artificial intelligence, presented last week and which Bettel will speak about on the occasion of the European commissioner for internal market Thierry Breton’s visit to Luxembourg next week. 

Finally, an awareness campaign on these themes--desired by 80% of respondents--will begin soon.

This article was originally published in French on Paperjam and has been translated and edited for Delano.