The recent BBC wire story, detailing accusations by a human rights group that Vietnam abducted Australian and Norwegian activists, brings into sharp relief a critical, often overlooked dimension of geopolitics in the age of advanced artificial intelligence: the asymmetrical power of information, or the lack thereof. While the immediate focus is on human rights and international law, the underlying mechanisms of how such narratives are constructed, disseminated, and believed—or disbelieved—are increasingly influenced by the very technologies I cover.
Vietnam claims the men entered the country illegally and were planning "terrorist activities." Conversely, the rights group alleges abduction. This is a classic "he-said-she-said" scenario, but the digital age imbues it with new complexities. In an ideal, transparent world, objective data – geolocation, communication logs, biometric scans, publicly available intelligence – could be fed into advanced analytical models to corroborate or refute claims. Machine learning algorithms, trained on vast datasets of human behavior, intent, and communication patterns, can, in theory, identify anomalies or corroborate sequences of events with high statistical confidence.
However, this is not an ideal world. In environments like Vietnam, where information flow is tightly controlled and independent reporting is severely curtailed, the input data for any such algorithmic analysis is heavily skewed, if it exists at all. The state controls the sensors, the networks, and often, the narrative. What appears in state media, what evidence is presented, and what counter-arguments are allowed to circulate are all filtered through a singular lens. When external observers, be they journalists or international bodies, attempt to apply analytical frameworks, they are met with a data vacuum.
This vacuum is where the true danger lies for human rights. Without verifiable, independent data, advanced AI systems, even those designed for truth-finding or anomaly detection, are hobbled. An AI model is only as good as the data it’s trained on and the data it receives for inference. If a state feeds its internal AI systems a curated dataset – one that consistently portrays activists as terrorists and state actions as justified – then the output of those systems will inevitably reinforce that narrative. Domestically, this can be used to justify repression and manufacture consent. Internationally, the lack of counter-evidence makes it challenging to dismantle the official story, even when it appears flimsy.
Consider the potential for deepfake technologies or highly sophisticated narrative generation AI. While these aren't explicitly mentioned in the BBC report, the underlying principle applies. If a state has the capability to generate compelling, albeit fabricated, digital "evidence"—whether it's doctored surveillance footage or AI-synthesized voice recordings—and can prevent contradictory evidence from emerging, then the burden of proof shifts dramatically. The absence of genuine data forces external actors into a position where they must either implicitly trust the state's narrative or make accusations based on ethical principle without definitive, verifiable evidence.
From a centrist perspective, the issue isn't about blindly trusting one side or the other, but about establishing robust, verifiable facts. My beat in AI and ML has taught me that the closer we get to general intelligence, the more critical the quality and diversity of our data inputs become. In contexts like the one described, the very mechanisms that could help adjudicate truth – advanced data analysis – are neutered by a lack of access and an abundance of state-controlled narrative. The implication for human rights is profound: without the ability to independently verify claims, the global community operates in a perpetual state of uncertainty, where justice becomes a function of political will rather than verifiable truth.
The lesson here is clear for policymakers and international bodies: as AI advances, so too must our mechanisms for establishing data provenance and independent verification, especially in regions where information is weaponized. Otherwise, we risk a future where digital authoritarianism, powered by sophisticated but biased AI, can operate with impunity, cloaking human rights abuses behind an impenetrable firewall of manipulated or absent data. The algorithmic shadow cast by controlled information flows threatens to obscure justice and empower narratives, regardless of their factual basis.