Đurašković: Media in Montenegro should adopt internal rules for the use of artificial intelligence

Media in Montenegro should adopt internal rules that will define for which tasks artificial intelligence (AI) can be used and in which situations its use is not acceptable, said project manager at the Media Union Dražen Đurašković.

He also pointed out that there is a danger that AI is not used primarily to support journalists, but rather as a means of producing a greater amount of content with a smaller number of employees, adding that this could lead to job losses.

Đurašković, who is also the author of the research "AI in the newsroom, the first changes in journalistic practice in Montenegro", said that AI is already present in Montenegrin newsrooms, but that its use is mostly taking place without clear rules.

He told the MINA agency that, although to a lesser extent, AI is also present in basic journalistic tasks such as composing journalistic questions, preparing drafts of texts, or suggesting headlines.

"What the research identifies is the intention of some employers to use employees' voices to train AI models, without employees being sufficiently informed about the scope and method of this use," added Đurašković.

He said that only 5,6 percent of respondents indicated that their media outlet had a policy governing the use of AI.

Đurašković said that about a quarter of those surveyed believe that when a text is processed by AI, it is no longer fully perceived as their own work.

"The survey also showed an extremely high level of agreement when it comes to the need for transparency: more than three-quarters of respondents believe that citizens have the right to know when AI has been used in the preparation of content, and almost half believe that non-transparent use can jeopardize trust in the media," said Đurašković.

According to Đurašković, digitalization has already changed working conditions in newsrooms, accelerated the production and distribution of content, increased the number of tasks that journalists perform daily, and changed the relationship between the media and the audience.

Asked what he sees as the biggest challenge when it comes to AI in journalism in Montenegro, Đurašković said that it is not the technology itself, but its uncontrolled and non-transparent application in newsrooms that are already working with a limited number of employees, under strong time pressure and without sufficient resources to verify information.

"In such an environment, there is a danger that AI is not used primarily as support for journalists, but as a means of producing a larger amount of content with a smaller number of employees," believes Đurašković.

This, he added, could lead to job cuts, increased workload for those who remain, increased monitoring of employee performance, and additional pressure to produce content faster.

"The consequence may be the publication of generic, superficial or insufficiently verified information," said Đurašković.

As he stated, the issue of introducing AI into newsrooms must therefore also be viewed as an issue of labor rights.

Đurašković said that the introduction of AI systems could affect the number of employees, job descriptions, workload, performance appraisal methods, professional autonomy and accountability of journalists.

"If journalists are expected to use AI, it is necessary to provide appropriate training, clearly define their obligations, and determine who bears responsibility for errors that occur when using these tools," said Đurašković.

According to him, journalists should not bear the consequences of decisions to introduce systems over which they had no control.

"A special problem is professional liability. AI can produce convincing but completely inaccurate information, a fabricated quote or a non-existent source. Therefore, a journalist and editor cannot simply shift the responsibility to technology," said Đurašković.

He added that every fact, piece of information, quote or source proposed by AI must be independently verified.

"At the same time, the employer must provide sufficient time and resources for such verification, because it is not possible to simultaneously demand from employees greater production speed and full responsibility for the quality of the content," said Đurašković.

He believes that journalists should not run away from AI, because it is a technology that will remain a part of media work.

"A complete rejection of its use would be neither realistic nor useful. Research has shown that journalists who use AI more frequently express a slightly lower level of uncertainty regarding its impact," said Đurašković.

He said that AI can be useful for transcription, translation, reviewing large numbers of documents, systematizing data, or language editing.

"However, it cannot replace fact-checking, talking to sources, going to the field, contextualization, and editorial judgment," Đurašković added.

He said that the journalist must know what the tool has done to the text, must be able to check the result and must have the final say.

Asked about the legal regulation of AI in the media, Đurašković said that the recently adopted revised Code of Journalists of Montenegro already contains provisions on the responsible use of artificial intelligence among its first guidelines.

As he said, this is significant because it shows that the profession has recognized the need to set basic ethical standards for itself before the issue is regulated in more detail by law.

"The Code insists on human verification and editorial responsibility for content created using AI, as well as transparency towards the audience when the use of these tools could mislead them," said Đurašković.

According to him, the bottom line is that responsibility cannot be transferred to technology.

"Regardless of how much AI is used, the published content must be backed by a journalist and editor who checked and approved it," Đurašković added.

This, he pointed out, is important because editorial control must remain in the hands of the journalistic profession.

Đurašković said that the state should prescribe the protection of fundamental rights, transparency and accountability of technology providers, but it should not determine how newsrooms will make editorial decisions or assess which way of using AI is professionally acceptable.

"Such issues should primarily be regulated at the professional level. Therefore, the new Code should be the starting point for each editorial office to clearly define the use of AI," said Đurašković.

He stated that the media should adopt specific internal rules that will define for which tasks AI can be used, in which situations its use is not acceptable, who checks the generated content, when the audience is informed, how sources and confidential data are protected, and who bears responsibility in the event of an error.

"Employees and their representatives should be included in their development, because they also regulate issues of work organization, professional autonomy, and labor rights," believes Đurašković.

Speaking about the state framework, he said that the development of the AI ​​Strategy of Montenegro for the period 2026–2030 is underway, while the integration process must also be harmonized with the European AI Act.

These documents, as he stated, can set a broader legal framework, but they cannot replace professional and editorial rules or take over editorial responsibility, which must remain in the hands of the journalistic profession.

Speaking about whether AI can replace journalists, Đurašković stated that this is not the most realistic scenario.

"The greater risk is that it will change the way work is organized and allow employers to produce a greater amount of content with fewer people," said Đurašković.

He said that AI cannot replace a journalist who investigates, asks uncomfortable questions, goes into the field and makes decisions in complex situations.

"That's why journalists shouldn't be afraid of technology itself, but they should be very careful about the way employers introduce it," believes Đurašković.

He said that the question is not just whether AI can perform a task, but whether it is used to improve the work of journalists or to reduce the number of employees, increase the norm and weaken the quality of journalism.

"This must be discussed openly at the professional level, through social dialogue, collective agreements and employee participation in decision-making," said Đurašković.