urges to distinguish administratively and criminally punishable violations / Article

In the Saeima, the discussion on the regulation of the use of artificial intelligence was started by the President Edgars Rinkēvičs, instigating amendments to the law.

“For the voter to be informed whether in a specific campaign material that he is currently reading, or if someone has called him or if he is participating in a video conference, he would know that he is interacting with a live and real candidate for the position of deputy or that he is facing an artificial intelligence system. For transparency and interaction between the voter and the deputy candidate to take place without any doubts”, the advisor to the President Edgars Stafeckis outlined the examples.

That’s why the law calls for a reference to those campaign materials created with the help of artificial intelligence. However, the representatives of the advertising industry pointed out in the discussion today that in that case it should be clearly defined in which cases there should be a reference to artificial intelligence tools, because

in the creation of advertising materials, they are widely used to enhance both text and images. Thus, reference to the use of these tools should be added to virtually every promotional material.

Anna Alyošina, head of the Political Planning and Communication Department of the Corruption Prevention and Combating Bureau, also agreed to this.

“I can clearly see that this would not be applicable to texts. There is an emphasis on images, audio and video. Not everything generated by artificial intelligence tools should be considered taggable or particularly noteworthy to the public. If we could narrow down and separate the wheat from the chaff , for which administrative responsibility arises and for which it does not. I think it would be a legitimate interest of the public to know about the use of artificial intelligence tools.”

At the same time, before the final reading in the Saeima, there are also amendments to the Criminal Law in connection with delaying elections by deception. There is still no consensus on whether this also includes so-called deep fakes with the use of artificial intelligence against someone. This would apply to video material in which a recognizable person states things they did not say. For example, the President with invitations to vote or not to vote for a specific party – such a dangerous example was given by Iveta Kažoka, head of the “Providus” center. His amendments to the Law on Pre-Election Campaigning called for a distinction to be made between cases where artificial intelligence tools were used in coordination with the candidate and when it was done maliciously.

“For example, when a candidate spreads a positive falsehood about himself, but the voter is not clearly told that it is not true information, but at the same time it is not dramatic information. Some voters would believe that it is a real person, some would believe that artificial intelligence was probably used. In such a case, there should be administrative responsibility. It is not the case that the image of the President is suddenly created, who says to vote for a particular candidate. This is much more harmful, and there should be criminal sanctions for this,” Kažoka explained.

In today’s discussion of the working group, in which no concrete decisions were made, the deputies emphasized that positive deep counterfeiting is also a dangerous and real possibility, which should be dealt with with the strictness of the law.

Anna Alyošina, an expert of the Bureau of Prevention and Combating of Corruption, however, explained that the discussions so far have already referred to the article of the Criminal Law, which stipulates responsibility for influencing the election if it was done fraudulently. Fraud also includes the so-called deep falsifications with the intention of undermining the reputation of an election participant or involving the image of a recognizable person in the creation of such materials.

Additional risks are also associated with possible foreign interference in Latvian elections, and this especially applies to the testing of artificial intelligence tools in our information space. Therefore, the idea supported by the National Security Council to create a competent institution – the Latvian National Center for Artificial Intelligence – is particularly welcome, emphasized Filips Rajevskis, head of the public relations agency “Mediju tilts”.

“What we need to focus on are the risks of blackmail and protection against foreign attacks on our election system. Latvia is a small country with well-developed IT. If someone experiments with these things, they will not do it in the USA, France or Germany, because it is expensive. We will be at the forefront and be experimented with these technologies because we are small!”

In the Saeima, the amendments to the Law on Pre-Elections still need to be considered in two readings, which are planned to be done by the end of the year, and the supplemented regulation would apply to the upcoming municipal elections next summer.