Author: Muhammad Rifky WildhaniEditor: Hosea Immanuel Latumahina

In 2024, Aliansi Jurnalis Independen (AJI) Indonesia released a briefing paper which includes a list of Indonesian news organizations that use artificial intelligence in their journalistic works. TVOne, for instance, is stated to use generative AI (GenAI) in their journalistic works, which led them to finally launch the first Indonesian news channel made with GenAI: tvOne.AI[1] on YouTube. Concurrently, generative artificial intelligence (GenAI) raises challenges in the field of journalism[2], even in Indonesia. Oftentimes, the concerns revolve around integrity, data privacy, lack of judgement from journalists as the sole creator, and even deterioration of trust[3]. On one hand, the use of GenAI on regular journalistic products might be acceptable under some regulations; on the other hand, we might want to take a look at the risks of GenAI use on investigative journalistic products.

What’s So Different about Investigative Journalism?

There are many different types of journalistic products, with no definitive accounts; but distinctly, sports journalism, popular journalism or lifestyle journalism, and investigative journalism seem to be the most produced genres of journalistic products. In any of these types, journalism aims to report trustworthy information to the public which is obtained by rigorous verification procedures[4]. This step has been the fundamental part of making journalistic products, at any rate. Say lifestyle journalism, for instance, are produced mostly to address public interests, such as entertainment, food, arts, and even gardening. For decades, as argued by Fürsich, we have experienced “the softening era” of journalism by massively producing lifestyle journalism[5].

On the other hand, Hunter & Hanson (2011) argued that investigative journalism is not irreconcilably and fundamentally distinct from other types of journalism, say lifestyle journalism for instance. Both types still demand the what, why, who, when, where, and how (5W+1H) framework. However in investigative journalism, the “why” question rather transforms to and emphasizes the “how” question. This difference varies investigative journalism to a distinct quality that reinforces emotional effect[6]. Additionally, investigative journalism involves a highly specific process of data-mining and highly dependent on skilled human work[7].

Furthermore, procedures in making investigative journalistic products also seem to differ from those of regular journalistic products; for it needs a further and longer process of data-collecting and verification[8]. In that regard, it is compulsory for journalists to collect not only undercover information, it is also mandatory for them to gather facts of incidents that would possibly harm the public if kept hidden, like corruption or any governmental violations that are often kept underneath public’s attention due to political circumstances[9]. Additionally, aside from its definition, investigative journalism poses more as a rigid methodology of conducting in-depth reporting that is grounded in the principle of public accountability[10].

Prior Development of AI in Journalism and Cases in Indonesian News Organizations

Prior to today’s type of GenAI machines in producing journalistic products, a United States-based news organization called the Associated Press back in 2014 started using automated and artificial writers. The new era of journalism has started ever since. Founder Robbie Allen stated that this system will help journalists to churn out articles with big amounts of data faster than ever. However, the founder admitted that there still remain challenges to journalism, as worried by many people[11]. One of the concerns revolves around contextualizing knowledge. Given the explanation of journalism’s nature to use 5W+1H framework in all outputs and emphasis on the “why” question, more pointedly in investigative journalism, it is risky for GenAI to take control in journalism.

Given the introduction, it is also no exception in Indonesian news organizations that they use AI, even specifically GenAI, during the process of their journalistic works. As mentioned previously, AJI Indonesia in 2024 had published a report consisting of a list of Indonesian news organizations that reportedly use AI in their journalistic works. Although it is not mentioned that all the Indonesian news organizations use GenAI specifically in their journalistic works, certain information about which news organizations that use GenAI is still provided: TVOne, Kompas Gramedia, and IDN Times. In April 2023, TVOne introduced a virtual presenter which is made with GenAI and launched tvOne.ai in 2024. On the other hand, Kompas Gramedia uses GenAI to create articles’ content, video, avatar, and even to make a conclusion. In addition, IDN Times had also launched “Timmy Si Robot”, a robot journalist made for producing 8 articles everyday and to cut down reading time to merely 30 seconds.

Speaking of Information Integrity and Ethical Concerns

Concerns may arise: why do we need a specific regulation to govern the use of GenAI in investigative journalism, especially in Indonesia? While GenAI gives advantage to journalists in analysing vast amounts of datasets, such as doing transcript, grammar-checking, and even reducing cost of production, it also poses a detriment. Generally speaking, one of the most debated harms GenAI does to investigative journalism is its human authority and information integrity of providing knowledge in trustworthy reports.

Speaking of GenAI’s epistemic disadvantage, Attard, et al. (2023) argued that GenAI cannot be used for fact-checking because it does not process contextual data like humans are able to[12]. To explain that condition, due to GenAI’s training system by large language models (LLMs), its systems are probabilistic[13]; they are potentially prone to errors, by which then they possibly produce misinformation, if it does not undergo human editorial judgement[14]. Moreover, using this model, GenAI tends to raise risks related to data manipulation and biases while automatically scraping and generation journalistic contents, which could reduce the integrity of the information[15]. Because, information integrity means that the information presented should be complete, verifiable, and understandable.

Additionally, there exists inescapable software of Google’s Gemini AI, which would always provide overview, generated from information available online. This would be concerning because the system would possibly generate information from sources that are actually presenting misinformation; thus, this activity would raise the possibility of AI by mass reproducing misinformation and deteriorate information integrity[16]. Further explanation of this case is provided by The New York Times which elaborates how synopsis by Google’s AI Overview are risky due to its possibility of gathering information that are linked to unsupporting websites; this condition is challenging to crosscheck AI Overview accuracy[17].

Concerns regarding ethics and legal risks, also arise and have always been in debate: data privacy, fairness, and accountability. These concerns raise due to the fact that not only GenAI, but AI systems in general collects data from available public sources to be extracted for and used for its system training[18]. Subjects’ private information used in the investigative journalistic works are also argued to be in the same concern[19]. As argued by Dodds et al. (2026), this raises a question regarding how do we know that a content (journalistic products, in this context) is not junk and verifiable?[20] Given the fact that investigative journalism often involves collecting underground kept-secret information, it is highly riskful to use GenAI in investigative journalism due to the methods of its training system.

GenAI on Journalism and How The Press Council Regulates It

Speaking of national scope, a regulation is established by the Press Council of Indonesia in 2025. Indonesia Law 5/2025 concerning AI use in journalistic works article 2 paragraph (2), strictly stated that, “the utilization of AI in journalistic works have to be under human control throughout the production.”[21] Moreover, this regulation is stringently grounded in Indonesia Journalism Code of Ethics/2006 article 3, in which it regulates that journalists need to closely cross-check information, balance coverage, prevent including personal opinion, and stand by the presumption of innocence. However, these regulations have not yet pointedly acknowledged the use of GenAI in investigative journalism.

Observing an international scale, there have been several regulations governing the use of GenAI in journalistic works, although not explicitly governing investigative journalism. In the United Kingdom, Independent Press Standards Organization (IPSO) governs that journalists are still in responsibility of their AI-generated and AI-provided journalistic works. In Vietnam, it is rigorously encouraged for journalists to review and be responsible for any journalistic works engaged with AI[22].

Conclusion

Cases of GenAI use in investigative journalism have been found, although not specifically conducted in Indonesia. Nevertheless, there are Indonesian news organizations that use GenAI in producing their journalistic works. These facts raise concern in epistemic and ethical aspects of journalism, pointedly investigative journalism that often associate with secret and riskful data-collecting from subjects. From time to time, ethical guidelines for journalists have always been built upon traditional belief that it is humans who judge journalistic works[23]. In Indonesia, the use of any type of AI is governed in Law 5/2025 by the Press Council concerning AI use. Given the arguments as mentioned many times previously about investigative journalism, it is endorsed that a specific regulation on GenAI use in investigative journalism have to take effect, respectively in Indonesia. This evaluated regulation hopefully does not take effect only on media organizations, but further, it takes to a legislative level by proposing information integrity and security to the Ministry of Communication and Digital Affairs of Indonesia.

1. AJI Indonesia. (2024). Artificial Intelligence (AI) dan Organisasi Berita di Indonesia: Pemetaan Pemanfaatan dan Rekomendasi untuk Stakeholder Media. AJI Indonesia. Accessed 19 September 2026 https://aji.or.id/system/files/2024-12/ajiartificial-intelligence-ai-dan-berita-dwibahasa.pdf. ↑
2. Ibid. ↑
3. Lewis, S., Guzman, A., Schmidt, T. et al. (2025). Generative AI and its disruptive challenge to journalism: an institutional analysis. Communication and Change 1, 9. https://doi.org/10.1007/s44382-025-00008-x. ↑
4. Harsono, Andreas. (2010). A9ama Saya Adalah Jurnalisme. Penerbit PT Kanisius. ↑
5. Fürsich, Elfriede. (2012). LIFESTYLE JOURNALISM AS POPULAR JOURNALISM. Journalism Practice, 6, 12-25. https://doi.org/10.1080/17512786.2011.622894. ↑
6. Hunter, M. L. & Hanson, Nils. (2011). What is Investigative Journalism, Story-Based Inquiry: A manual for investigative journalists, (pp. 8-10). ↑
7. Wellbrock, Christian-Mathias. (2024). Generative AI and Investigative Journalism—An Application of Baumol’s Cost Disease Across Mass Communication Sectors. Emerging Media, 2(4) 663–672. https://doi.org/10.1177/27523543241299631. ↑
8. Kurnia, S. S., Kuswarno, E., Maryani, E., & Firmansyah. (2021). The characteristics of investigative news organizations in Indonesia between 2010-2012. Heliyon, 7(2). https://doi.org/https://doi.org/10.1016/j.heliyon.2021.e06276. ↑
9. Ibid. ↑
10. Ibrahim, Al Ali. (2025). Investigative Journalism in the Age of Artificial Intelligence: from Evidence to Impact. Independent Academic Publication. ↑
11. Jenkin, Matthew. (2016). Written out of the story: the robots capable of making the news. Accessed in September 2026 from https://www.theguardian.com/small-business-network/2016/jul/22/written-out-of-story-robots-capable-making-the-news.
12. Attard, M., Davis, M., & Main, L. (2023). Gen AI and Journalism. Centre for Media Transition. https://doi.org/10.6084/m9.figshare.24751881. ↑
13. Ibid. ↑
14. Ashuri, T., Zalmanson, L., & Goldstein, D. (2026). Journalistic Epistemic Authority in the Age of AI: The Incorporation of Generative AI in Newsrooms. Digital Journalism. https://doi.org/10.1080/21670811.2026.264042. ↑
15. Mendiburu, M., Sayad, Alexandre Le Voci, & Durán, Tomás. (2026). Information Integrity and Media and Information Literacy as enhancers of the Right to Access Information. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000397850. ↑
16. Dodds, T., Zamith, R., & Lewis, S. C. (2026). The AI turn in journalism: Disruption, adaptation, and democratic futures. Journalism, 27(3) 530–544. https://doi.org/10.1177/14648849251343518. ↑
17. The New York Times. (2026). How accurate are Google’s A.I. overviews?. Accessed in September 2026 from https://www.nytimes.com/2026/04/07/technology/google-ai-overviews-accuracy.html. ↑
18. Attard et al. (2023). ↑
19. Ibid. ↑
20. Ibid. ↑
21. Dewan Pers. (2025). Pedoman Penggunaan Kecerdasan Buatan dalam Karya Jurnalistik. Accessed in September 2026, from https://jdih.dewanpers.or.id/dokumen/peraturan/peraturan-dewan-pers-nomor-1peraturan-dpi2025-tentang-pedoman-penggunaan-kecerdasan-buatan-da. ↑
22. Luat Vietnam. (2026). AI-generated press content must be labelled under new rules. Accessed in September 2026, from https://english.luatvietnam.vn/legal-updates/ai-generated-press-content-must-be-labelled-under-new-rules-892-110271-article.html. ↑
23. Lewis et al. (2025). ↑