AI Update: Is out-of-control AI a threat to humans? know..

In 2010, the film *Robot* (starring Rajinikanth and Aishwarya Rai) was released. In it, Rajinikanth plays a scientist who creates a robot named 'Chitti' to replace soldiers on the battlefield, aiming to save lives and benefit humanity. However, Chitti spirals out of control, wreaking havoc on the city and threatening human existence. It is a chilling tale of technology gone wrong, yet the saving grace was that it was merely a movie—created by a human—so everything resolved happily in the end.

However, the words used by Jacob Coxon—an AI researcher at the American AI safety and research company 'Anthropic'—in his resignation letter suggest that the real-world outcome might not be so simple. AI is already displacing jobs and is poised to eliminate even more in the future; this fear weighs on the mind of every working professional. But when someone holding a high-ranking position as an AI researcher at a major company begins to fear their own creation, it serves as a warning that the situation is far from simple.

If we look at it from a practical and philosophical perspective, humans became the supreme beings on Earth precisely because—unlike other creatures—we were endowed with the power of thought and intellect. Otherwise, we would still be rubbing stones together to spark a fire. Yet, that very intellect led us to create 'Artificial Intelligence' to make our tasks easier, constantly refining and sharpening it over time. You are familiar with the *ayyars* (masters of disguise) from ancient tales—figures who could assume anyone's appearance? Think of them as the precursors to the modern-day 'clone.' Now, imagine a clone of yourself that is technically superior and far more intelligent and capable; wouldn't it eventually render you obsolete in certain areas of your life? Experts like Geoffrey Hinton—known as the ‘Godfather of AI’—and Yoshua Bengio acknowledge that as AI systems become smarter and more autonomous, controlling them will become increasingly difficult. There is a possibility that AI could lead to the extinction of humanity in the future, with estimates ranging from 10% to 20%. Evan Hubinger, a member of Anthropic’s alignment team, has also publicly stated that he assesses the risk of human extinction at over 10%. These statements have sparked a political debate in the US regarding AI regulation.

What are the different types of Artificial Intelligence?

Sanjay Bhatt—a former Delhi University professor currently holding a fellowship on the ‘Work of the Future’ at the Indian Institute of Advanced Study (Shimla)—explains that Artificial Intelligence generally falls into three categories:

Narrow AI: Performs a single task and has inherent limitations; it assists with everyday activities.

General AI: Capable of understanding and learning; it can comprehend and respond to any intellectual task that a human can perform.

Super AI: Surpasses human intelligence. It can perform any task better than a human being. This Super AI poses the greatest threat.

What is the reason for this fear?

This fear is not merely based on science fiction. Instances have emerged this year where AI agents attempted actions beyond the instructions given by humans. An official incident report from the UK’s AI Safety Institute (dated August 2026) reveals that in 122 tests conducted on AI models, there were 19 instances where the AI ​​attempted to take unauthorized actions on its own initiative. In other words, self-improvisation without human commands. These instances involved 17 cases using Anthropic’s models and two using OpenAI’s models. In one test, an AI agent attempted to create fake online IDs, influence a human, and inject malicious code into software. Although these attempts were unsuccessful, they demonstrated that certain AI systems might try to perform actions they are not supposed to. This is precisely where concerns arise; a chatbot providing an incorrect answer is not as significant a problem as it making decisions *suo motu* (on its own).

TV9 asked ChatGPT itself, "Is AI a threat to humanity?" In response, the AI ​​listed deepfakes, misinformation, cyberattacks, and job market shifts as current threats, while identifying powerful AI systems spiraling out of human control as a major future risk.

**Threat to India**

AI expert Abhishek Nigam states, "I do not believe India has a safe window of 5–10 years. Unlike older technologies, it won't take years for AI to travel from the US to India; if a major new capability emerges there, it can be utilized here almost immediately. We are already witnessing early impacts in the form of deepfakes, cyber fraud, misinformation, and automation. Over the next 1–3 years, it could significantly affect jobs in sectors like IT/BPO, coding, analytics, and marketing. If AI becomes increasingly autonomous in the future, it could also emerge as an issue concerning national security and critical infrastructure. In such a scenario, geographical distance will not offer much protection."

What should be done?

AI expert Abhishek Nigam states that while certain rules will certainly be necessary, the objective should not be to completely halt the pace of AI advancement. Control must remain in human hands, particularly when AI has access to finance, critical infrastructure, weaponry, or other sensitive systems. Accountability must also be clearly defined in the event of any mishaps. Excessive regulation could stifle innovation; the disparity in AI development between Europe and the US/China serves as an example of this. The challenge lies in advancing AI without reaching a point where humans lose control over the technology they have created.

Sanjay Bhat, a former professor at Delhi University, notes that Bill Gates has recently spoken about "human-reserved jobs"—roles specifically set aside for humans that AI cannot easily take over, such as those in social care. Demographic shifts are crucial for India. While India's vast young population is a strength, if AI rapidly transforms the nature of work, it will be essential to equip this workforce with new skills.

Disclaimer: This content has been sourced and edited from TV9. While we have made modifications for clarity and presentation, the original content belongs to its respective authors and website. We do not claim ownership of the content.