The conventional narrative about an AI existential threat often focuses on a "superintelligence" that "wakes up" and decides to wipe out humanity, or a rogue AI that misinterprets a human-given goal with catastrophic results (e.g., "end all human suffering" becomes "end all humans").

However, another perspective, which many experts find more plausible and immediate, focuses on how AI could lead to an existential crisis without becoming a conscious, malevolent entity. This perspective views AI as a powerful tool that, in the hands of humans, could dismantle the very foundations of our society and governance, leading to a "slow-moving catastrophe."

Here are some key points of this alternative view:

1. The Erosion of Shared Reality

• Disinformation at Scale: AI-powered tools, especially large language models and deepfake technology, can generate and disseminate hyper-realistic fake content—articles, videos, and audio—at an unprecedented scale and speed. This could lead to a world where it's impossible to distinguish truth from fiction.
• Fractured Society: When a shared understanding of reality breaks down, it becomes impossible for societies to function. Democratic processes, public trust, and social cohesion depend on a common set of facts. AI-driven disinformation could create "echo chambers on steroids," where groups of people inhabit completely different realities, making compromise and collective action on global problems (like climate change, pandemics, or nuclear proliferation) impossible.

2. The Amplification of Human Stupidity and Malice

• Automated Warfare: AI systems could be integrated into military decision-making, accelerating the pace of conflict to a point where human oversight is obsolete. This raises the risk of accidental nuclear war or a global conflict spiraling out of control due to an algorithmic error or a "flash war" triggered by AI-driven systems on opposing sides.
• Bioengineering and Pandemics: A highly capable AI could serve as an "accelerant" for malicious actors, dramatically lowering the barrier to entry for developing and releasing novel biological or chemical weapons. It could provide the knowledge, simulations, and even the necessary designs to a lone terrorist or rogue state, making it far more difficult to contain.
• Cybersecurity Catastrophe: An AI could find and exploit vulnerabilities in a nation's critical infrastructure (power grids, financial systems, communication networks) at a speed and scale that is humanly impossible to defend against. This could lead to a total systems collapse of modern civilization, plunging the world into chaos.

Dear Adnan Majeed

Thank you for this valuable contribution

Kind regards

Mohamed

Dear Mohamed el Nawawy

The real risk arises from AI becoming a powerful tool that is simple to use and easily programmed by many who have access to enough computing resources and data enabling them to use this technology in a way which contradicts with ethical values.

https://www.linkedin.com/pulse/two-distinct-views-ais-potential-existential-threat-yuri-7j7je/

Regards,

Shafagat

Dear Shafagat Mahmudova

I'd like to emphasize that democratization of powerful AI tools highly contributes to the risk to humanity posed by AI, which is caused by simplicity of AI programming and availability of computing resources and data as described in my discussion.

Regards,

Mohamed