1. The Warning from Jakarta: When Machines Feed on Bad Data
In a recent statement that has reverberated through Indonesian media circles, Deputy Minister of Communication and Digital Affairs Nezar Patria raised an alarm that many journalists have been quietly whispering about for months: artificial intelligence is changing the way news is made, and not always for the better. Speaking on Thursday, Nezar pointed to a subtle but dangerous phenomenon called “data poisoning.” It sounds like something from a cyber-thriller, but it is a very real problem. AI systems do not invent information out of thin air; they learn and generate responses based on the data they are fed. If that data is corrupted, outdated, biased, or simply false, the AI will confidently produce misinformation. And when journalists use AI to write articles without going into the field, that misinformation slips into news reports, reaching millions of readers who trust what they see. The deputy minister’s warning is not an abstract philosophical debate. It is a practical and urgent issue for newsrooms across Indonesia and indeed the world. The media industry is at a crossroads. On one hand, AI offers incredible efficiency: it can summarize reports, draft headlines, translate languages, and scan vast amounts of information in seconds. On the other hand, that efficiency can come at the cost of accuracy, the very foundation of journalism. Nezar’s comments highlight a growing concern among media watchdogs and press councils: if journalists become overly reliant on AI, they may stop doing the essential work that defines the profession—going out, observing events, talking to people, and verifying facts. The result could be a news ecosystem where readers cannot distinguish between what is real and what is a plausible-sounding fabrication generated by an algorithm. This is not about being anti-technology. It is about being pro-truth.
2. Data Poisoning: The Hidden Flaw in Artificial Intelligence
To understand why data poisoning is such a serious threat, it helps to think of AI like a student who reads thousands of books and then writes essays based on what they have absorbed. If the books contain errors, the essays will repeat those errors, often with great confidence. In the world of artificial intelligence, data is everything. Large language models and generative AI tools are trained on massive datasets scraped from the internet, archives, social media, and published works. Much of that data is reliable, but a significant portion is not. It may contain outdated statistics, unverified claims, hoaxes, or deliberate disinformation planted by bad actors. When AI draws on this inconsistent pool of information, it can produce outputs that sound authoritative but are actually wrong. Nezar explained that AI cannot function without data, and that data can be flawed. The term “data poisoning” refers to the manipulation or contamination of training data, whether intentionally or accidentally, which causes AI to generate incorrect or harmful results. In a newsroom setting, this could mean an AI tool confidently reporting that a government official made a statement they never made, or citing a study that does not exist, or combining unrelated facts into a false narrative. Journalists who use AI as a research assistant or ghostwriter may not always recognize these errors, especially when time is short and the pressure to publish is high. The danger is compounded by the fact that AI writing tends to be smooth, fluent, and persuasive. It does not hesitate or show doubt. It presents its output as established fact. For a reader scrolling through a news feed, there is no obvious red flag. The misinformation is baked into the article, and it spreads with alarming speed across social media, messaging apps, and online platforms. Nezar’s warning is therefore not just about the quality of journalism; it is about the integrity of public knowledge in a digital age.
3. The Erosion of Traditional Journalism: Observation, Verification, and Truth
The deputy minister also touched on a more philosophical yet equally practical concern: the erosion of traditional journalistic methods. For more than a century, the gold standard of journalism has been direct observation and verification. A journalist covering a protest does so by being there, watching the crowd, listening to the speakers, counting the numbers, and interviewing participants and witnesses. A journalist reporting on a government policy does so by reading the actual document, asking experts, and seeking responses from different sides. This process is time-consuming and difficult, but it is precisely what separates journalism from rumor, opinion, and propaganda. Nezar recalled that journalists used to go outside, observe events, interview sources, and verify information before turning it into a news story. Today, AI can streamline the entire process by gathering information from various sources without any verification. A journalist can type a prompt into an AI tool and receive a draft article in seconds, complete with quotes, statistics, and contextual background. But where did those quotes come from? Were the statistics checked? Is the background accurate? In many cases, the journalist does not know. The AI does not reveal its sources, and if challenged, it cannot defend its claims. This represents a serious challenge to journalism, not only because it threatens the business model of media organizations, but because it attacks the most fundamental principle of the profession. Nezar reminded his audience that the number one code of ethics for journalist organizations worldwide is to report the truth. If journalists abandon the discipline of verification and delegate their judgment to algorithms, they risk becoming mere conduits for machine-generated content rather than trusted chroniclers of human events. The business pressure is real: newsrooms are shrinking, deadlines are tighter, and AI offers a way to produce more content with fewer resources. But the cost of that convenience is high. Once a news organization loses its reputation for accuracy, it is very difficult to regain. Readers may turn to other sources, or worse, they may stop trusting the news altogether.
4. Government and Regulatory Response: Guardrails for AI in the Newsroom
Recognizing the gravity of the situation, the Indonesian Ministry of Communication and Digital Affairs has begun to issue guidelines on the ethical use of artificial intelligence. The ministry has released a circular aimed at protecting the public from misinformation, disinformation, and hoaxes that can be amplified by AI tools. The message to journalists is clear: you may use AI, but you must do so responsibly. The government has directed journalists and media companies to align their AI usage with the guidelines formulated by the Press Council, the independent body that oversees press freedom and ethics in Indonesia. This is a significant step because it acknowledges that AI is not going away. It is already embedded in news production workflows, from automated transcription to content recommendation algorithms. The question is not whether to use AI, but how to use it without compromising the principles of accuracy, fairness, and accountability. The circular likely includes recommendations such as always verifying AI-generated information through primary sources, clearly labeling content produced or assisted by AI when appropriate, and ensuring that a human editor reviews every article before publication. It also places responsibility on AI users rather than on the technology itself. Nezar emphasized that the government urges AI users to ensure the accuracy of the information generated by artificial intelligence. This is a subtle but important shift. Instead of trying to ban the technology or regulate it into submission, the government is asking the people who wield it to exercise caution and good judgment. This approach aligns with broader global trends. Many countries are debating AI regulation, but the most effective frameworks are those that focus not only on the developers of AI systems but also on the users and the context in which AI is deployed. In journalism, that context is sacred. The goal, as Nezar put it, is to ensure that the public does not fall victim to misinformation, disinformation, or hoaxes. That is a public interest objective that transcends politics and business.
5. The Broader Battle for Trust in a Misinformation Age
The implications of Nezar’s warning extend far beyond the newsroom. In a world where social media algorithms are already accused of amplifying division and falsehood, the entry of generative AI into the information ecosystem makes the problem even more complex. Bad actors can use AI to create fake news articles, forged images, synthetic audio, and convincing deepfake videos. But perhaps the more insidious threat is the subtle erosion of trust. When the public begins to suspect that news articles might be generated by machines, that photos might be altered, and that quotes might be fabricated, their confidence in all media diminishes. This creates a vacuum that rumors, conspiracy theories, and propaganda can fill. It also makes it easier for politicians or other powerful figures to dismiss legitimate journalism as “fake news.” Nezar’s comments suggest that the government is aware of this risk and feels a responsibility to protect the integrity of public discourse. However, the fight against misinformation cannot be won by regulations alone. It requires a collective effort involving journalists, technology companies, educators, and readers. Journalists must re-commit to the craft of verification. Technology companies must design AI tools that are transparent about their limitations and data sources. Educators must teach media literacy skills so that citizens can critically evaluate what they see online. And readers must be willing to pause before sharing an article, to ask who created it, and whether the claims are backed by evidence. In this environment, the role of the journalist is more important than ever. Human judgment, emotional intelligence, and ethical reasoning are things that AI cannot replace. A machine can describe an event, but it cannot understand the grief of a mother who lost her child in a disaster. A machine can summarize a court ruling, but it cannot perceive the injustice embedded in a legal loophole. These are human insights that require presence, empathy, and experience.
6. Finding Balance: AI as a Tool, Not a Replacement
The challenge, then, is not to reject AI entirely but to integrate it wisely into the practice of journalism. AI can be an incredibly powerful tool if used correctly. It can help journalists process large datasets, identify trends, translate interviews, transcribe speeches, and even provide background research. It can free up time for the solitary and essential work of field reporting, source building, and deep analysis. But it must never become a substitute for direct observation and verification. Nezar Patria’s message is a reminder that technology should serve humanity, not the other way around. The media industry is already facing an existential crisis. Advertising revenues have collapsed, traditional business models are struggling, and audiences are fragmented across countless digital platforms. The temptation to use AI as a shortcut to cut costs is understandable. But the long-term survival of journalism depends on something that cannot be automated: trust. A reliable and credible press is one of the essential pillars of a functioning democracy. Without it, citizens cannot make informed decisions, governments cannot be held to account, and the public sphere becomes a battleground of competing fictions. As Indonesia moves forward with its digital transformation, the words of the Deputy Minister should serve as a guiding principle for journalists and media organizations everywhere. Embrace innovation, but do not abandon the fundamentals. Use algorithms to assist in research, but never allow them to dictate the truth. Protect the professional standards that define journalism, and remember that every article, every headline, and every broadcast carries the weight of the public’s trust. In the end, data poisoning can only harm us if we willingly drink from the poisoned well. The antidote is human vigilance, ethical judgment, and a renewed commitment to the ancient craft of bearing witness. That is not just a lesson for Indonesian journalism. It is a lesson for the world.