The rapid proliferation of digital technologies is reshaping the global security landscape. What was once an interconnected, but relatively limited, digital world, has altered into a complex interconnected ecosystem where dangers and threats lurk at every step. The combination of Artificial Intelligence (AI) and cyber threats has emerged as one of the most critical concerns of the modern world.

While AI has changed from being merely an operational efficiency mechanism into an absolute infrastructure mandate, cyber threats are becoming more sophisticated, harder to detect and more difficult to mitigate. Today, AI-powered malware can adapt and evolve in response to changes in its environment, making it harder for traditional anti-virus software to detect. The global interconnectedness of digital systems has also raised the ante as far as cyber attacks are concerned. Malicious autonomous agents are today well positioned to undermine the legacy of “Zero Trust” protocols, aggravating insider threat vectors.

Meanwhile, as seen in the war in Ukraine, newer technologies are demonstrating the potential to undermine conventional military calculus through asymmetric impact. The rapid proliferation of digital technologies and systems across every discipline, cyber breaches and malicious activities are impacting every level of security. Many governments facing sophisticated cyber adversaries are drawing up a security architecture which involves a blend of innovation, policy coherence and strategic realism, but still lag behind.

A new disruptive era

AI, meanwhile, is reshaping every consequential domain of human activity. AI’s Large Language Models (LLM) are intruding into many business enterprises, without either proper training or evolution of the human interface. As ‘Agentic’ AI — as distinct from Generative AI — becomes a force to reckon with, problems attached to AI implementation are set to become even more complex. The catalyst for such concerns is the arrival of the latest AI systems, notably Mythos, developed by United States firm Anthropic. According to experts it demonstrates an uncanny ability to detect ‘Zero day’ vulnerabilities across major operating systems. What is evident is that AI is beginning to test the limits of legacy systems.

AI’s overarching influence makes it essential to understand the risks and threats it poses, especially when combined with cyber capabilities, where the danger grows exponentially. AI has already transformed modern warfare, as seen in the Russia-Ukraine conflict and the U.S./Israel-Iran war. AI-powered smart systems can detect, track, and predict missile trajectories with remarkable accuracy. AI can also provide different kinds of intelligence from photos, text, radio, and electromagnetic signals to deny adversaries a tactical advantage. Munitions, such as systems dubbed “Goalkeeper” and “Whiplash” can independently identify and strike targets. Yet, nations, particularly the U.S., have not yet demonstrated a level of sagacity on how best to wield this power and in a manner that does not endanger human civilisation.

These developments are fuelling concerns about AI’s future as increasingly capable models and robots replace human roles. LLMs and autonomous agents are already performing complex tasks once reserved for people. The World Economic Forum warns that while AI will strengthen cyber defences, it will also enable more sophisticated automated attacks, potentially rendering current “Zero Trust” protocols ineffective. This raises concerns that hard power may increasingly be shaped by nations and organisations at the forefront of AI.

Companies in the West such as Anthropic and Palantir are behaving as if they own the world of tomorrow, since they hold the key to the world’s mightiest technology. Anthropic has reached a stage today where with Mythos, it creates the spectre of all-encompassing power residing in a single instrument that is not subject to any rules. China, perhaps, has its own version of Anthropic or something resembling it, but this is of little comfort to the rest of the world. What is becoming clearly evident is that an attacker with the latest version poses a grave threat to not merely one country but an entire civilisation.

An AI ‘power shift’

As AI technologies become more rampant and uncontrollable, existing rules concerning cyber security may require to be completely rewritten. Many experts warn that as “Agentic Autonomous Operations” become even more sophisticated, the world could reach an inflection point that has grave forebodings for the world. Whether Mythos is the harbinger of this is yet to be determined, but it is generally acknowledged that the newer generations of AI machines provide opportunities for ‘rogue groups’ to demonstrate capabilities that were once associated with only nation states. As Agentic AI begins to displace Generative AI, enabling it to act more autonomously, threat levels are certain to go up still further. AI has, hence, the potential to become the dominant source of military and economic power, even as the rules to govern AI are still in a fledgling state. The bugle call has already been sounded by the pontiff of Anthropic, the AI firm that makes models such as Mythos, that it has the potential to become the dominant source of military and economic power as civilisation advances.

Notwithstanding the god-like properties of existing modern AI models, experts warn that they are often subject to “hallucinations”, thus giving scope for distortions, producing misleading reactions. Judgments in such cases tend, quite often, to be subjective. Even otherwise, biases are known to creep into decisions dictated by AI, enhancing the danger posed by unrestricted reliance on AI. Furthermore, experts warn that it is vitally important to guard against such ‘algorithmic radicalisation’, and ensure that AI platforms do not push opinion makers towards extreme views.

A future that needs oversight, wisdom

The world, thus, appears to have reached a tipping point in AI progress and development. Already, some companies that are leading the AI charge are accusing others of stealing their technologies and expanding on them to their advantage. The latest is a charge made by the U.S. against China, viz., that China’s ‘Moonshot’ AI stole from Anthropic’s most sophisticated AI Large Language Model ‘Fable’, to help create its latest AI version. Predictably, China has rejected this charge, but it reflects the intensity of the competition between the U.S. and China in this area. If Anthropic’s claim is true — that its latest model has surpassed “all but the most skilled humans” in finding and exploiting security vulnerabilities in everything from popular operating systems to cryptographic software — the world could face a new kind of apocalypse.

The world has steadily progressed from abiding concerns of stagnant economic growth to doomsday technologies, and this merits a great deal of introspection. Altruism is not enough to decide on how to handle a technology that is about to eat up the world. There are several aspects of advanced AI that clearly demand more introspection, but this is not evident. As AI increasingly enters into the arena of defence sensing and warning, there is a need not only to keep machines under human control but also to ensure that AI-dictated warnings do not outrun proper understanding and reality.

Those steeped in defence aspects already know the gap that exists between warning and understanding of what might happen, and how critical this is in strategic terms. High-powered AI systems may often prove unreliable as they may demonstrate a bias towards national security given that algorithms often display such biases. The world, hence, needs to reach an understanding on how best to deal with this new technology which can, and could, end or make the future of the world. What we have seen in Ukraine and West Asia is not even the tip of this iceberg. Better to “cross the river by feeling the stones” is a better maxim than ‘blind prejudice or unadulterated endorsement of a new technology’.

M.K. Narayanan is a former Director, Intelligence Bureau, a former National Security Adviser, and a former Governor of West Bengal

Published - August 04, 2026 12:16 am IST