Our ability to think is under serious threat from Artificial Intelligence (AI). We are not grasping the severity of this threat quickly enough, nor foreseeing its consequences. The rapid proliferation of AI means it is urgent to act now to protect this most basic of human abilities. The consequences of failing to act could be widespread and irreversible.
The explosion in the uptake of AI models is accompanied by alarm bells rung from many corners of society on unprecedented risks: job displacement, more severe cyber-attacks and warfare, increased prevalence of online abuse, violence and misinformation, and cognitive decline.
Evidence is emerging about cognitive decline; however, the most important point is missed. AI violates the fundamental human right to think. All other human rights are predicated on this fundamental right – without the ability to think, we cannot protect ourselves against violence of all other human rights.
The United Nations General Assembly adopted the Universal Declaration of Human Rights on 10 December 1948, in the aftermath of World War 2. Its 30 Articles set out the fundamental rights and freedoms of every human being to live and work in security, free from discrimination, violence and unfair treatment.
In the sixty years since their adoption, these human rights have been enshrined in national law and, despite backsliding in many countries, still serve as the foundational pillars of rights to which all human beings are entitled, and by which all must abide.
Article I says, ‘All human beings are born free and equal in dignity and rights. They are endowed with reason and conscience and should act towards one another in a spirit of brotherhood.’ We cannot reason without thinking.
There are many ways in which thinking is defined and analysed in academic literature. In this article, I equate thinking to cognition. Its Latin root provides clues to its definition: cognicioun, ‘ability to comprehend, mental act or process of knowing.’ Further, the term cognition can be defined as ‘the activities and processes concerned with the acquisition, storage, retrieval and processing of information – regardless of whether these processes are explicit or conscious.’ (Byrne et al. 2019)
If cognition is the ability to comprehend or learn information, how does AI impact this? For example, to write this article, I searched Google Scholar for academic articles on thinking. I read and processed information from various papers, and, having compared different definitions, I made a judgement to select the one I thought would be most readily understood by a wide audience. This felt hard.
I was reading papers with different academic definitions, based on a wide range of psychology and multidisciplinary theories. The answer wasn’t easy to obtain. I had to spend a few hours learning and then deciding – but that was time spent learning and thinking.
If I had asked one of the large language models (LLMs) to find a definition of thinking, I would have found an answer within seconds. There is no need to spend a few hours acquiring, storing, and processing information on the definition of thinking. We clearly don’t need to spend time learning and thinking about every aspect of our work and lives, but it is a phenomenon we are now forced to confront as AI gains the ability to think for us at an extraordinary scale.
The academics at MIT have shown what many of us intuitively know. They tested the impact of LLM-assisted essay writing on 54 participants by monitoring the cognitive performance of students who completed essays using LLMs, a search engine and brain-only. They found that ‘brain-only participants exhibited the strongest, most distributed networks [brain connectivity]; Search Engine users showed moderate engagement; and LLM users displayed the weakest connectivity.’
Further, ‘over four months, LLM users consistently underperformed at neural, linguistic, and behavioural levels. These results raise concerns about the long-term educational implications of LLM reliance and underscore the need for deeper inquiry into AI’s role in learning.’
Much has been discussed in the public domain about the harm related to social-media driven fake news and misinformation. AI and LLMs worsen this risk in multiple ways.
They act as an information gatekeeper– mediating and structuring content from various sources, which can cause cognitive and other vulnerabilities. (Potnis et al. 2024) One of the many problems with this is that AI-generated content can include ‘substantial gender and racial biases.’(Fang at al. 2024) In subtle and noticeable ways, content generated by AI continues to be prone to hallucinations, which makes the fact-checking and systems for identifying errors even more important. (Spatharioti et al.2024) The compounding effects of these issues on the information we access and what we think is frightening. There is an urgent need for comprehensive, ethical, governance and mitigation strategies.
Thinking is a complex phenomenon, but let’s, for argument’s sake, assume that it is shaped by the ability to learn. What is the ability to learn?
In 2010, the OECD published a book titled The Nature of Learning. It collates a wide range of research sources on learning environments, cognitive perspectives on learning, the crucial role of motivation and emotion in learning, as well as many other important areas. Cognitive perspectives on learning are summarised through ten ‘cornerstones,’ one of which is that learning requires time and effort.
AI not only reduces but often eliminates the time and effort required to access information, which consequently reduces the need or ability to learn.
Without any guardrails to protect our ability to think in the age of AI, what changes to human cognition are underway? In high-pressurised corporate environments, how are employees’ tasks and learning experiences changing? What about the risks?
In July 2026, the Financial Times (FT) published an article on how Thought Leadership reports produced by one of the big four were ‘riddled with fake footnotes, misattributed claims and unverifiable information,’ because of ‘exhorting staff to use the technology themselves to speed up and improve their work.’ Besides the embarrassment, what real-world consequences are playing out as a result of AI hallucinations seeping into public reports? In the case of the big four, their reports carry weight – they shape views and opinions which influence decision-making.
How are students using AI in their education? Particularly those under pressure from exams. How is their ability to learn changing as a result? Is there a risk that AI deepens the socio-economic divide? For example, poor university students who must earn an income to support their studies, with less time to learn than wealthier peers, may well feel more pressure to use AI to meet deadlines.
The World Economic Forum’s Future of Jobs report (WEF, 2025) identifies analytical thinking as the number one skill in 2025. Creative thinking is at number four on the list and systems thinking at number 12. Critical thinking and human judgement are frequently cited by employers as increasingly important key skills. Research finds a ‘significant negative correlation between frequent AI tool usage and critical thinking abilities, mediated by increased cognitive offloading’. Younger users who use AI more frequently display lower critical thinking skills. (Gerlich, 2025) The paradox is that our ability to develop critical thinking skills is under threat, whilst those very skills are increasing in demand as AI use proliferates.
Further, the ability to learn the truth is predicated on preserving the time it takes to learn and think. As Yuval Noah Harari says, ‘The truth is costly, if you want to know the truth about anything, you need to spend a lot of time and energy and effort looking for it – and factchecking and investigating.’ Indeed, the truth can be difficult and painful. However, if we agree that it is a non-negotiable bedrock of the human condition as we know it and the social fabric of societies, then we must do all we can to preserve it.
Besides time, access to information and teaching in all its forms is an important ‘cornerstone.’ The widespread reports of AI hallucination and misinformation further diminish the learning experience. (Boretti, 2026)
Who will protect the time and effort it takes for us to learn? To understand the truth, whether in education or at work. Can we rely on non-state actors to do this? If the answer is no, then we must look to public institutions to act in the public good.
The Universal Declaration of Human Rights was developed after the event. It is not only that it is far better to prevent than cure, but in this case human existence as we have known it for centuries may well depend on prevention. The most important prevention may be to protect our ability to think by the rule of law. To do this, it is vital to protect the right to think as a fundamental human right.
Note: No AI model was used to write this article.
-30-
Copyright©Madras Courier, All Rights Reserved. You may share using our article tools. Please don't cut articles from madrascourier.com and redistribute by email, post to the web, mobile phone or social media.Please send in your feed back and comments to [email protected]
