How Indonesia’s media landscape is dealing with AI

AI made its way into Indonesian media organisations in 2023. A year later, outlets began using AI across production workflows. They apply it to editing, automated tagging (software-generated labels for organising content), voice generation and even the creation of on-screen avatars.

A 2024 briefing paper by the Alliance of Independent Journalists (AJI) described how industry leaders tested AI for content distribution and technical assistance as well as creative applications such as headline suggestions. The paper also highlighted risks: disinformation, algorithmic bias and unresolved questions about copyright protection for AI-generated or AI-assisted material.

By 2026, the use of AI had shifted from experimentation to routine practice, according to a BBC Media Action study titled “Understanding the Use of AI in Indonesian Newsrooms”. Based on a survey of 212 journalists, the study reports that 75 % are using AI tools as part of their daily work.

The study also shows which tools dominate. ChatGPT leads by a wide margin: 86 % of respondents rely on it, followed by Gemini (63 %) and DeepSeek (12 %). For design and other creative tasks, Canva ranks highest (32 %). Respondents also mentioned Adobe Podcast or CapCut for audio and video editing. Journalists remain more hesitant about using AI for verification. Only 28 % turn to it for factchecking, though nearly half of that group relies on it every day.

Overall attitudes are ambivalent rather than enthusiastic or dismissive. Around 70 % describe AI as an opportunity, but 45 % of the study’s participants simultaneously perceive it as a potential threat. More than half associate AI with benefits such as faster workflows, greater creative capacity and improved data analysis. At the same time, 30 % warn of harms, including weaker journalistic values, fewer job opportunities and heightened risks of plagiarism and content duplication.

Crawling and zero-click searches

Indonesia’s media industry, like media elsewhere, faces disruptive change as AI rapidly develops. Generative AI has begun to affect both the business models of media companies and the long-term sustainability of newsrooms.

Wahyu Dhyatmika, chairperson of the Indonesian Cyber Media Association (AMSI), describes two recent shifts that, in his view, have reshaped the landscape. First, generative AI tools have proliferated, notably ChatGPT and Google AI Overview, which he links to “zero-click searches” in which users get answers on a platform without visiting the publisher’s site. Outlets that depend on organic traffic (visitors who get to a website via unpaid search results) have seen visits fall, sometimes by more than 50 %, he says, as audiences increasingly turn to AI platforms for condensed summaries. These summaries often draw on reporting produced by the very outlets losing traffic.

Secondly, Dhyatmika points out that automated bots that scan websites (“crawlers”) have also surged in number on news sites. He says more bots now extract and record journalistic content without licence agreements or payments to publishers, which is creating a serious challenge for the industry. According to Dhyatmika, fewer than five percent of AMSI’s roughly 500 members have put technical controls in place, such as machine-readable instructions for bots or whitelist/blacklist protocols to manage crawler access. AMSI is therefore developing infrastructure based on OpenMind AI to monitor crawling activity and pursue compensation from AI systems that mine journalistic content. The organisation is currently piloting the system with three members.

Backlash over AI errors

The use of AI in newsrooms has already sparked some controversy in Indonesia, with several outlets facing public backlash over errors linked to AI-generated content.

CNN Indonesia came under criticism last year after publishing an article about US President Donald Trump that still contained AI-generated recommendations. The error quickly spread on social media and raised questions about basic editorial procedures.

In January, the newspaper Radar Tulungagung was accused of circulating false information about an alleged promise by President Prabowo Subianto to appoint new civil servants. The article also used an illustration sourced from Gemini.

Together, these incidents highlight a key risk of AI in journalism: the pursuit of speed and automation can compromise accuracy and, with it, credibility. As Indonesian journalists continue to experiment, they face the same practical challenge as their counterparts around the world: How can AI be integrated into workflows without compromising journalistic standards?

AI pioneer

Since the launch of ChatGPT in November 2022, Kompas.com – one of Indonesia’s first online news outlets – has positioned itself as an active AI pioneer in the country’s media landscape. “We see that society and technology are moving toward AI and civilisation itself will follow. That’s why we explore AI,” says Johanes Heru Margianto, managing editor of Kompas.com, which is part of the large Kompas Gramedia Group. Kompas.com gives journalists and content creators broad leeway to test AI and use it throughout the workflow, from pre-production to distribution.

The outlet has also integrated AI into its content management system (CMS), the software used to produce and publish articles. In this setup, AI helps detect typos and suggests alternative story angles and narrative options. At the same time, Margianto says the newsroom is cautious about using AI as an information source because it can be inaccurate and therefore misleading, especially when staff lack the relevant factchecking skills.

That concern prompted Kompas Gramedia’s media unit to formalise rules. In October 2023, Margianto and 11 colleagues drafted the KG Media Guidelines on AI Utilization. The document sets out practical requirements for AI use and specifies legal and disciplinary consequences for violations, ranging from written warnings to termination.

The KG Media Guidelines frame AI as a supporting tool that must remain under human oversight. Margianto argues that journalists may use AI to compare and develop ideas, but they should avoid fully delegating content creation to AI, as this weakens the authenticity of journalistic work. Even for routine content – such as weather forecasts, prayer schedules or sports fixtures – editors at Kompas.com still curate AI-produced material.

Indonesia’s Press Council later published its own guidelines in January 2025. They emphasise that newsrooms must align AI use with journalistic ethics and that journalists remain ultimately responsible for verifying any AI-assisted content.

Economic pressure

Economics has become a second, closely linked challenge. As AI quickly spreads, it reshapes the media industry’s basic assumptions and raises questions about outlets’ long-term competitiveness and relevance. Kompas.com has responded by blocking AI bots from automatically extracting and saving (“scraping”) its articles and has chosen instead to sell its journalism directly to AI platforms.

Margianto links the shift to changes in audience engagement. “Globally, the largest traffic source has always been Google. That’s why SEO became so important,” he says, referring to search engine optimisation – tactics that help content rank highly in search results. With AI-generated answers and summaries, however, Google-driven traffic is declining worldwide. In Indonesia, he adds, virtually all online outlets are now seeing shrinking traffic.

At Kompas.com, Margianto estimates the decline at roughly 20 %, which he considers less severe than the drops other outlets have reported. Still, he argues that AI does more than change consumption habits: it undermines the very mechanics of the media economy. “The core of the media business is the crowd. That’s what gets monetised. When the crowd disappears, the business foundation collapses,” he says. For him, the fundamental problem goes beyond monetisation: the media must redefine their role as their traditional functions – informing and educating the public – are increasingly being taken over by social media and AI systems.

AI for in-depth reporting: Zona Utara’s pivot

Other companies are taking a different route. Zona Utara, a North Sulawesi media outlet, has experimented with AI in newsroom operations for the past three years. But instead of competing in the high-speed daily news cycle, it now prioritises long-form and investigative reporting.
Founder and editor-in-chief Ronny Buol says the shift reflects changing audience behaviour. “People don’t open Google anymore. They go straight to AI. So why should we keep producing daily news?” he asks. Zona Utara still publishes daily stories when major events occur, but it no longer organises its work around routine updates.

The new editorial focus does not mean the outlet has abandoned AI. Over the past year, Zona Utara has advanced by developing its own tools. Buol argues that using general-purpose systems for news invites “hallucinations” and misinformation. “That doubles our workload because we have to verify everything,” he says.

So, Zona Utara built its own AI agents into its CMS. The agents follow strict prompts that mirror newsroom standards: they must cover the 5W+1H (who, what, when, where, why and how), include verification steps and provide a clear byline. Buol says he and his colleagues taught themselves to build these tools despite not having an IT team.

Even as it shifts toward investigations, the outlet still publishes “evergreen” content – also with the help of AI. Buol cited ship schedules as an example: Zona Utara takes data from Indonesia’s national cargo and passenger shipping company, converts the datasets into articles and produces interactive maps and visuals.

To protect quality, the newsroom introduced internal AI guidelines in late 2024. Reporters may not use generic AI tools, both to reduce bias and to protect sensitive internal data. The rules also require transparency: any AI-assisted content must carry a disclaimer.

The pivot is paying off. Since mid-2025, readers have spent more time on the site. “We have loyal audiences for in-depth coverage,” Buol says, adding, “maybe because AI doesn’t provide investigative reports.”

Anastasya Andriarti is a journalist based in Indonesia who also works as a lecturer in journalism and a researcher in media and communication studies. 
sinaudata@gmail.com