Humanity once knew when to say « No ». In 1997, when the world discovered that the Dolly sheep had been cloned after its birth the previous year, governments, ethicists and the public reacted with concern, not complacency. In a few weeks, US President Bill Clinton ordered federal agencies not to finance human cloning and called for its prohibition, while ethics and law catch up with science (Clinton 1997). In November 1997, UNESCO adopted the Universal Declaration on the Human Genome and Human Rights, denouncing human cloning for reproductive purposes, and in 1998 the Council of Europe amended its Convention on Biomedicine to ban it altogether (UNESCO 1997). Even as technology opened up new perspectives, governments and public opinion agreed on a simple principle: certain limits must not be crossed (Center for Public Integrity 2022; CIRM 2011).
Today, the world faces an even more destabilizing border: the fusion of artificial intelligence with autonomous weapons and space defence systems. Instead of stopping this development, we are doing exactly the opposite. Instead of a moratorium, we are witnessing an acceleration. Instead of serving the interests of humanity, we strive to enrich a handful of capitalists.
For example, Palantir. Born in the wake of September 11 with the initial support of In-Q-Tel, the CIA venture capital branch, Palantir forged a reputation by supporting US intelligence services and special operations in the « war against terrorism », especially during high-profile manhunts such as the decade-long Osama bin Laden hunt. Its data integration platforms allowed analysts to merge intercepted communications, financial leads, travel records and field reports into targeting files, discreetly integrating the company into the post-11 September war architecture.
What began as a tool to combat terrorism is today the central nervous system of the algorithmic war. What was originally a post-11 September deficit analysis start-up, which was only worth a few billion at the time of its direct stock exchange in 2020, is now a military AI giant valued at hundreds of billions.. Palantir currently manufactures billionaires while its software helps to wage wars that ordinary people pay for with their blood.
Two decades later, under Trump's second mandate, Palantir was repositioned at the heart of a new defence system: an AI complex that openly supports and directs the development of AI for military purposes in the context of a renewed confrontation between major powers in Washington (Carnegie Endowment 2024). Its software now provides advanced targeting, logistics and battlefield management for US forces and their allies, particularly in the war between Russia and Ukraine, where the MetaConstellation and Gotham Palantir platforms exploit satellite images, drone flows and electromagnetic intelligence to generate lists of targets to be eliminated in near real time (In Depth News 2025). Ukrainian officers describe targeting cycles reduced from several hours to a few minutes, with hundreds of Russian positions hit with precision artillery and rockets guided by analyses provided by Palantir (In Deep News 2025; CNA 2023). For investors, this is a proof of principle; For civilians on the ground, it is the algorithmic industrialization of death.
Even more worrying is the role of Palantir in the emerging project of « Golden Dome » A space and air defence architecture that integrates AI, surveillance constellations and interception systems into a global and permanent combat network (CIGI 2024). Presented as a shield, it is actually a gigantic « Star of Death » in the sky, a combat platform capable of tracking and hitting targets located thousands of kilometers from the safety of the sky. With the help of constellations of commercial and military satellites, high-altitude drones and ground sensors, di-IA pipelines can identify patterns, report activities « Suspects » and triggering weapons systems with minimal human intervention (Tech Policy Press 2025; CNA 2023). In theory, it promises precision; In practice, it allows large-scale killing of innocent people from orbit or high altitude, based on models and metadata that no court or victim will ever see. Golden Dome is not just defense; It is the standardization of a permanent global targeting infrastructure.
All this happens as the AI itself accelerates beyond the human rhythm. The CEO of NVIDIA, Jensen Huangrecently warned that China « a few nanoseconds of delay » on the US in terms of AI and advanced chips, signaling an arms race where the gap between competitors is measured in fractions of microseconds. When Palantir-type systems are integrated into Golden Dome-type architectures, decision-making loops are further shortened: satellites observe, the AI class, networks decide, and autonomous or semi-autonomous systems strike, most certainly faster than diplomats cannot write discussion points, not to mention de-escalation plans (CIGI 2024; Diplomacy.edu 2025). The same logic of profit that once led Palantir to monetize the fight against terrorism is now driving this company and its peers to monetize strategic deterrence, space domination and autonomous strike capabilities. Each crisis becomes a real-life demonstration; every climb, a sales argument.
If Dolly marked a moment when humanity took the time to think and opted for restraint, Palantir's journey – from the track of Ben Laden in the shadow to the implementation of targeting systems based on AI and space combat worthy of « Golden Dome » – shows a civilization that heads down to disaster. There is still no binding global treaty banning fully autonomous weapons; EU pioneering AI law explicitly excludes military AI from its scope; UN discussions on autonomous lethal weapons are continuing without reaching an agreement, even though systems are being tested in Ukraine, Libya, Gaza and elsewhere (Carnegie Endowment 2024; Human Rights Watch 2025; ScienceDirect 2025). In the worst case, a type architecture « Golden Dome » related to command systems enhanced by the AI wrongly interprets a satellite manoeuvre, a naval incident or a missile test as an attack, and the machines trigger an escalation before humans can intervene.
In a world where interstate rivalries are only a few « nanoseconds of deviation », particularly in the development of AI-powered intelligence capable of making critical decisions that can cost human lives, where companies benefit from each new automation layer, and where civilians only appear as data points in a targeting graph, the margin of error is disappearing. Contrary to what happened with Dolly, we didn't draw a red line, and it may already be too late. If we persist in not pulling the hand brake, the combination of Palantir-type war platforms and Golden Dome-type space defences could well ensure that ordinary human beings become the first, and last, victims of an AI-driven calculation error, too fast to be stopped and too devastating to survive.
Raja Shuresh
First Counsellor – Embassy of Malaysia in Senegal
