By 2035, Our Daily Lives May Be Dramatically Transformed by AI
By 2035, our daily lives may be dramatically transformed by AI. Office work led entirely by AI — presentations, emails, data analysis, and report creation, all intellectual tasks generated by AI. In meetings, everyone turns to AI for answers and pastes the response. In entertainment, chart-topping songs and blockbuster films churned out by AI in just a few days. In education, memorizing knowledge has become a relic of the past; students now focus on learning how to operate AI tools specialized for specific domains. This is the arrival of a "prompt economy" — where everything is available at the cost of a single prompt. Not long ago this sounded like science fiction, but in 2026, with AI technology being embedded into every piece of technology we use, it is no longer an unrealistic vision of the future.
But this rapid penetration of AI confronts us with a serious question: will we stop relying on our own brains? Will we gradually lose our problem-solving ability? Will AI "soften" our minds and make autonomous thinking impossible? In short: is AI making us less intelligent?
This article focuses specifically on the potential risks of consumer AI overuse, exploring its impact and what we should do about it.
- From GPS to LLMs: The Impact of Technology Dependence on Brain Function and the Reality of "Cognitive Offloading"
- Algorithmic Complacency and "Model Collapse": AI's Distortion of the Information Environment and a Warning for Decision-Making
- Avoiding the Thought Shutdown: The Importance of Critical Thinking in the AI Era and Balancing Productivity
- Summary: "Thinking Ability" as a Compass for the AI Era
Looking to optimize community management?
We have prepared materials on BASE best practices and success stories.
From GPS to LLMs: The Impact of Technology Dependence on Brain Function and the Reality of "Cognitive Offloading"
To think carefully about AI's impact on our thinking ability, we first need to understand the fundamental mechanism by which the human brain adapts to — and becomes dependent on — technology. A relatable example is GPS navigation systems like Google Maps, which many people use daily. A 2020 study showed that frequent GPS use can reduce spatial memory capacity, while also delivering economic benefits. Interestingly, subjects in this study who showed memory deterioration didn't perceive themselves as having poor directional sense, despite what the data showed. This is a good example of how convenience often comes with invisible costs. And GPS systems are relatively simple applications — they're not strictly AI. Yet even their "overuse" can damage our memory capacity. AI — particularly recent large language models (LLMs) — may have a far more complex and powerful influence.
Going back roughly five years before the GPS study, David Raffo, a professor at Portland State University, was deeply concerned about the quality of his students' papers. The logical structure was weak, academic depth was lacking, and faculty broadly sensed a decline in student motivation to learn. Then, in the middle of the pandemic, something unexpected happened: some students' writing quality dramatically improved. Raffo sensed something wrong in this unnaturally rapid progression — a slight improvement would have made sense, but the leap was too large, beyond the range of normal human growth. When he directly questioned the students, he discovered they had been using AI writing tools. His response: "I realized it wasn't that their writing had improved — the tool had improved the writing."
The word "skill" carries important meaning here. Raffo didn't categorically dismiss AI use, but he noted its effects were "a double-edged sword." He acknowledged the benefits — "AI enables us to complete work quickly and efficiently through rapid collection and organization of information across written communication, design development, and suggestions for tackling difficult problems" — but added this warning: "Our mental and cognitive functions are like muscles that need to be used regularly to remain strong and active. In engaging with technology currently available, having the self-discipline necessary to remain mentally strong and active is a challenge that would be difficult for anyone but an exceptional individual." This is a profoundly suggestive observation. Could chronic AI overuse cause a kind of "mental atrophy" — cognitive deterioration from lack of cognitive exercise? And as Raffo points out, resisting the temptation to take the easy path and continuing to make the effort to think for oneself has become a genuinely difficult challenge in the modern era.
This discussion has critical implications for our long-term health as well. Alzheimer's researcher Dr. Ann McKee, speaking on the "Diary of a CEO" podcast, noted that remaining mentally active and maintaining autonomy are important habits for dementia prevention. When examining the brains of people who lived to 85, about half showed the pathological signs of Alzheimer's disease — but not all of them developed symptoms. High cognitive reserve — that is, high cognitive capacity — strengthens the brain's resilience and can suppress the onset of symptoms even when pathology is present. By actively using the brain and continuing to challenge it, even if part of the brain has sustained damage, other pathways can compensate for the function, potentially allowing a person to avoid experiencing symptoms.
With this in mind, using AI to perform simple recall or observation on our behalf is not advisable. For example, asking AI to identify book titles on a shelf behind you — as seen in Google's Gemini demonstrations — is essentially handing over your own attentional effort to AI. We need to be much more careful about how these small accumulated actions affect our cognitive capacity.
History shows that the impact of calculators on basic math skills and the potential for autocorrect to undermine students' punctuation and spelling ability have long been areas of study. These tools assist specific skills, but the latest language models like ChatGPT, Llama, and Grok have the potential to go beyond assistance and substitute for our "thinking itself." This is genuinely a step into unknown territory.
It's widely predicted that routine clerical work — data entry, bookkeeping, customer service — will be replaced by AI, and that change has already begun, producing some concerning results.
AI's factual errors are well-known, but recent research shows that increasing dependence on AI triggers a phenomenon called "cognitive offloading" — the tendency to reduce the mental effort required to execute tasks by using external tools and systems. Researchers studied how AI use affects critical thinking skills across more than 600 diverse participants. They found that "frequent AI users showed a stronger tendency to outsource mental tasks, relying on technology rather than engaging in independent critical thinking when problem-solving and decision-making. Over time, participants who relied heavily on AI tools showed a decline in their ability to critically evaluate information and draw nuanced conclusions."
This cognitive offloading is beginning to affect not just individuals but social systems as well. A shocking example emerged in the legal system. In 2026, Detroit police, with only blurry surveillance footage as a clue in a liquor store robbery case, turned to the facial recognition system of DataWorks Plus — a crime database management company founded 25 years ago that now uses AI to support law enforcement. AI analysis surfaced Porsha Woodruff as a match to a 2015 mugshot. She had a prior arrest record for an expired license, but when police went to arrest her, she was eight months pregnant — in no condition to commit a violent crime like robbery. Nevertheless, police arrested the wrong person based solely on DataWorks' AI analysis, and she subsequently suffered from dehydration and labor complications. The case was ultimately dropped for lack of evidence, but this was not the first time Detroit police had succumbed to cognitive offloading. They currently face three lawsuits related to wrongful arrests based on DataWorks' AI, and similar cases continue to emerge.
From the outside, police negligence may seem obvious. But the problem runs deeper: this technology is sold and used as a "reliable alternative." People over-trust AI's capabilities because it makes life easier. And as the GPS research showed, once something becomes part of daily routine, the adverse effects become harder to recognize. The allure of convenience is that powerful.
Concrete examples of cognitive offloading are visible on social media platforms like X (formerly Twitter). There is a never-ending stream of users asking Grok AI to explain even extremely simple posts. More and more people are abandoning the act of thinking for themselves and uncritically accepting the answers AI provides. Whether this represents progress or regression is a judgment left to each individual.
Algorithmic Complacency and "Model Collapse": AI's Distortion of the Information Environment and a Warning for Decision-Making
Asking AI a simple question on social media might look, on the surface, like a rational time-saving behavior. But the tendency toward overuse — relying on AI even for things you could easily understand by thinking it through yourself — raises a warning flag about our thinking habits. Even for people who believe they don't use AI this way, there is an important lesson to take from this phenomenon: the extent to which we are unconsciously delegating decision-making to algorithms every day.
Instagram, Facebook, Twitter, TikTok, YouTube — the content we see on these platforms is personalized by algorithms. You may have discovered this article through YouTube's recommendation algorithm. The problem is that people are gradually surrendering their agency without realizing it. The more dependent we become on algorithms, the less frequently we ask ourselves what we actually want to see or know. Ultimately, what we see is determined not by us but by the algorithm.
Alec Watson of the Technology Connections YouTube channel named this phenomenon "algorithmic complacency." He warns about how infrequently we decide for ourselves what to look at and do on the internet. "What I find particularly novel and concerning is that I'm starting to see evidence of people who actually prefer to have a computer program decide what they see when they log on, even when they know alternatives exist," he notes.
The contrast with internet experiences from decades ago is stark. The internet was once accessed primarily through desktop web browsers. Google was a search engine for finding websites, and sites you liked were saved to your own bookmarks for future visits. Internet navigation and curation were highly manual — users were in control.
But generations coming of age in the 2020s are said to trust algorithms more than other humans. This tendency is also reflected in the reality that students who used AI to skip basic learning skills during and after the pandemic are now bringing those habits into the workplace — and many are increasingly relying on additional AI tools to compensate for their skill gaps. Is this "smart working"? Or is it gradually eroding long-term mental resilience?
Of course, for simple, repetitive tasks where AI doesn't make mistakes, time savings are real. But if people continue delegating all of their thinking to AI, they will end up thinking almost nothing for themselves. In that sense, AI has the potential to dull us and degrade our thinking ability — particularly when it begins to substitute for critical thinking itself.
Since the mid-1990s, the internet has led us into the information age. Search engines, social media, and YouTube accelerated that trajectory. Now, AI is synthesizing that vast information into "knowledge," and some say we've entered the "knowledge age." In theory it sounds great — but if that "knowledge" is flawed and most people can't detect the flaws, our grip on reality begins to slip.
This problem surfaced when Google introduced "AI Overviews" — an AI-generated summary block appearing at the top of search results. The launch was disastrous. Misinformation like "Obama was the first Muslim commander-in-chief," "snakes are mammals," and "eating a stone a day is good for your health" was generated, exposing AI's glaring weaknesses. In a few years the technology may be nearly flawless, but at present, hallucinations and misinformation based on inappropriate sources remain fundamental problems, and trust in AI has been damaged.
The deeper problem is that people accept these false pieces of information uncritically and spread them on other platforms as if they were facts. This is the core issue. AI is fundamentally different from the other technologies mentioned above — because AI still makes many mistakes. One survey found 70% of people trust AI news summaries, and 36% believe AI models provide factually accurate answers. Yet a BBC study last year found that more than half of AI-generated summaries from ChatGPT, AI agents, Gemini, and Perplexity had "significant problems." Even simply asking ChatGPT to make a sentence more polished can distort the meaning of the original text — and most people don't notice the change.
A still more serious problem is the phenomenon known as "model collapse." In early 2025, Oxford University researchers investigated what happens when AI reads and rewrites content generated by AI repeatedly. After just two prompts, output quality deteriorated significantly, and by the ninth prompt, output had become completely incoherent. As AI reuses data it generated itself — often inaccurately — as training data, the divergence from reality grows with each cycle, and the quality of knowledge degrades.
The study's lead researcher, Dr. Ilia Shumailov, explained: "What is remarkable about model collapse is how fast it happens and how hard it is to notice. First it affects minority data — data that is not sufficiently represented. Then it affects the diversity of output, and variance decreases. Sometimes there is a slight improvement in performance on majority data, which hides the performance degradation on minority data. Model collapse can have serious consequences."
But the most alarming finding comes from a separate study by Amazon Web Services (AWS) researchers. It suggests that approximately 60% of internet content this year (at the time of the study) may have been generated or translated by AI. If this figure reflects reality, AI technology is causing the internet to collapse in on itself — proliferating increasingly inaccurate information with each cycle. The trajectory leads to one of two outcomes: AI technology advances rapidly enough to avert the worst-case scenario, or the internet becomes saturated with inaccurate, incomprehensible "AI slop." This proliferation of AI-generated content lends credence to the "dead internet theory" — the claim that most content on the internet has already been replaced by bot- and AI-generated material. AI may eventually become an excellent distiller of knowledge, but in its current early stage, there is a possibility that it is pushing us backward.
Avoiding the Thought Shutdown: The Importance of Critical Thinking in the AI Era and Balancing Productivity
The rapid evolution of AI and its penetration into society raises many concerns. But before casting AI as a pure villain or falling prey to excessive fears of losing free will, it is important to engage calmly with its true nature. Large language models differ in character from GPS and spell-check — but they also share something with those tools: they are "tools," devices for performing specific functions. Fear of automation is not new. History shows that technological innovation has always caused temporary anxiety. But when properly implemented, automation has ultimately improved our productivity.
A good example is VisiCalc, developed in 1979 by Dan Bricklin and Bob Frankston — the first serious spreadsheet software for personal computers, designed to dramatically accelerate spreadsheet processing. At the time, changing a single number in a large spreadsheet meant redoing all the calculations and correcting them by hand. VisiCalc became the first "killer app" and was a major reason people bought personal computers. Initially, experienced computer enthusiasts didn't understand its value — "you can do this with BASIC," they said. But when accountants saw VisiCalc in action, the reaction was dramatic. Bricklin recalled one accountant "shaking, saying 'this is what I do all day.'" Of course, this automation did not eliminate the accounting profession. Rather, those who understood and could leverage the new tools drove change and raised the industry's overall productivity.
AI language models, handled responsibly, have the same potential to extend our capabilities as VisiCalc did. The crucial shift is treating AI not as something that thinks for you, but as a companion that assists your thinking. And AI's answers must always be received with a grain of salt. As Thomas G. Dietterich, professor of computer science at Oregon State University, points out: "We tend to interpret and use large language models as if they were knowledge bases, but in fact they are not knowledge bases — they are statistical models of knowledge bases." Simply put, LLMs are designed to give lengthy answers to questions even when they have no useful information to provide. They will never say "I don't know."
Dietterich further emphasizes the importance of systems having "a model of their own competence — knowing precisely which questions they can answer and which they should not attempt to answer" — and argues this thinking should be extended to LLMs. Neural network technology, by its nature of learning self-representation, faces a fundamental limitation: it can only express things it has encountered in some form of variation in the past.
The sensationalism around AI has certainly gone too far in some respects, but some of the concerns are legitimate. AI undoubtedly has a big future, but "the moment" has not yet arrived. One common problem identified in research findings from universities, think tanks, and research institutes is that people "trust" AI as their primary information source too much. Think back to a newspaper photo from 1988 showing teachers protesting the use of calculators in elementary schools. Their demand was not the complete removal of calculators from schools — it was opposition to "early introduction." The goal was to ensure young children could first solidly learn mathematical concepts themselves. We need to take the same approach with AI.
AI should be a "tool" for getting things done and moving forward more efficiently. But its use should not cost us the capacity to understand and resolve complex problems ourselves. No matter how sophisticated AI becomes, humans and their critical thinking will remain indispensable. We humans have real experience and a complex, nuanced understanding of the world around us. Until AI dominance actually materializes, humans should respect and cherish the capacity to think for themselves. René Descartes' first philosophical principle — "Cogito, ergo sum" (I think, therefore I am) — is as widely known as it is for a reason. Thinking is the foundational activity that makes us human.
To maximize the benefits of AI while avoiding the risk of thought shutdown, it helps to keep the following in mind:
Treat AI as a co-pilot: Rather than delegating all your thinking to AI, use it as a tool that supports parts of the thinking process — brainstorming, assisting with information gathering, proofreading writing. Final judgments and deep reflection must always be done by your own mind.
Maintain a critical eye: Never accept AI-generated information uncritically; build the habit of fact-checking. Especially for information involved in important decisions or domains requiring specialized knowledge, cross-referencing multiple reliable sources is essential. AI generates responses based on statistical patterns, so it may occasionally include misinformation or biased perspectives.
Maintain and improve foundational abilities: Even if AI handles your writing and calculations, don't neglect efforts to maintain and improve basic reading comprehension, writing ability, logical reasoning, and arithmetic. These foundational abilities are the basis for evaluating AI output and using it appropriately. Regularly accumulating experience thinking and solving problems independently contributes to maintaining cognitive capacity.
The same perspective is required in the workplace. Particularly among younger employees (ages 22 to 39), there is a growing movement to leverage AI tools to reduce workload. Some surveys show more than 90% of younger employees using two or more AI tools per week. Being freed from inefficient tasks — like spending more than 30 minutes finding the right tone for an email — is a genuine benefit. AI can improve productivity and contribute to business growth, more efficient management, and smoother team communication. But what this article has repeatedly warned against is the "overuse" of LLMs — relying on "AI slop" and abandoning the act of thinking with one's own gray matter. If the generation entering the workforce now fails to pay attention, they risk falling into excessive AI dependence, and like muscles atrophying around a broken bone, their creative thinking capacity could wither.
Summary: "Thinking Ability" as a Compass for the AI Era
AI technology holds the potential to bring revolutionary change to our lives and how we work. Its benefits are immeasurable, and when used appropriately, it will contribute enormously to productivity improvement and the creation of new value. But as this article has detailed, the "overuse" of consumer-facing AI and large language models carries the risk of seriously impairing our cognitive abilities and critical thinking. "Cognitive offloading" — outsourcing our thinking — "algorithmic complacency" — surrendering our agency — and the degradation of knowledge quality through "model collapse" and misinformation are real and present challenges we must be alert to.
Like GPS, calculators, and VisiCalc before it, AI is a powerful "tool." What matters is not being dominated by the tool, but using it proactively on our own terms. Position AI as an assistant to your thinking, or as a co-pilot — always maintaining a critical perspective, with final judgments made by your own mind. And beyond relying on AI, the key to navigating the AI era wisely is continuing the effort to maintain and improve your own thinking capacity, reading comprehension, and writing ability.
No matter how AI evolves, the human capacity to understand complex reality, make nuanced judgments, and conduct ethical reflection remains indispensable. As Descartes saw, "thinking" is the essence of being human. Precisely because we now hold this powerful tool in our hands, we must value and train our own thinking ability more than ever before. In a society of coexistence with AI, thinking ability is the compass that will open the path to the future.
Reference: https://www.youtube.com/watch?v=iqVhUX4Vel8
Streamline Event Management with AI | TIMEWELL Base
Struggling with large-scale event operations?
TIMEWELL Base is an AI-powered event management platform.
Track Record
- Adventure World: Managed Dream Day with 4,272 participants
- TechGALA 2026: Centrally managed 110 side events
Key Features
| Feature | Result |
|---|---|
| AI page generation | Event page completed in 30 seconds |
| Low-cost payments | 4.8% fee (industry-leading low rate) |
| Community features | 65% continue engaging after events |
Feel free to reach out for a consultation on streamlining your event operations.