Generative AI (i.e., systems that produce content in response to prompts) has advanced rapidly over the past year. ChatGPT’s latest updates, including the release of GPT-4.5 and its new image generator, have taken the internet by storm. The creative possibilities are striking, just look at the flood of Ghibli-style images users have shared online.
At the same time, Google has released Gemini 2.5, a major leap in its own AI development, with impressive reasoning and computation capabilities. The momentum is undeniable, and with each release, the capabilities of these models push closer to what some are calling the AGI (Artificial General Intelligence) moment.
AGI, as Shannon Vallor (2024) defines it, is “a machine with ‘human level’ competence across a wide and open-ended range of cognitively demanding activities.” (p. 22). Whether the development of AGI is good or bad is a complex debate for another time.
What matters right now, especially for those of us in education and academic research, is this: as AI tools become more integrated into how we think, write, and learn, we need to ask hard questions about how they’re being used.
Are they undermining our critical cognitive abilities or are they empowering us to think more deeply by freeing up mental bandwidth for creativity, analysis, and reflection? The answer, I believe, doesn’t lie in the tools themselves but in the intentions, habits, and awareness we bring to them.
To even begin to approach this question meaningfully, we need to start with a simple but vital distinction: Is AI assisting our thinking or is it replacing it?
This distinction helps us reflect on how we use AI and on whether it’s helping us grow intellectually or quietly outsourcing the very work that defines learning and research.
There’s a tendency to draw comparisons between this moment and past technological breakthroughs, like the calculator, the internet, or spellcheck. Yes, we adapted then. But this is different.
For the first time in history, we’ve created a system that simulates aspects of human cognition. These models don’t just follow instructions, they respond in language, offer suggestions, reason through dilemmas, and even reflect our tone and style. In some ways, they seem to think.
And their “thinking” power is immense. These models have consumed more text than any human could read in hundreds of lifetimes. That’s the kind of cognitive reach we’re dealing with. It’s not surprising, then, that researchers have raised concerns about the potential impact of over-reliance on AI particularly when it comes to critical thinking and deep learning (e.g., Gerlich, 2025; Smith & Funk, 2024)
This body of research is important. But it should be used to inform better practices, not justify blanket bans or reactive policies. As I see it, the problem isn’t AI. It’s how we choose to use it.
If you’re using AI to help clarify your ideas, refine your writing, or explore new angles, that’s responsible use. If you’re asking it to do the thinking for you, generate your arguments, or write your paper wholesale, that’s something else entirely.
We should be aiming to use AI as a thinking partner, not a thinking replacement. Ethan Mollick (2024) calls this co-intelligence, and I think that’s the right frame. You bring your insight, experience, and curiosity to the table, AI helps you sharpen and expand it. That’s the relationship we should be building.
In the book I’m currently writing on the use of AI in academic research, I emphasize this distinction throughout. Generative AI can be a powerful ally, especially when viewed as a research assistant: helpful, supportive, but never in charge. It can help researchers explore new directions, refine complex ideas, and push through blocks. But it should never be the generator of your ideas.
What I often tell researchers is this: develop your argument first. Get your ideas down. Then use AI to refine, restructure, polish. Don’t start with the tool, start with your thinking.
I argue that one of the most liberating aspects of using generative AI in research and writing is this: we no longer have to carry the full cognitive load of ‘languaging‘ our thoughts.
For many of us, especially non-native English speakers, translating ideas into academic language has always been a significant hurdle. It drains mental energy and often distracts us from the heart of the work: the thinking itself.
You might not realize just how many cognitive resources are spent not on generating ideas, but on packaging them in the right words, syntax, and structure. For some, that effort is so taxing that the original insight gets diluted or even lost.
But now, with generative AI, we have an “external brain”, a language partner that can take over part of that burden. This doesn’t mean we should hand over our ideas. It means we can now focus more freely on the originality, coherence, and depth of our thinking, rather than getting bogged down by grammar or phrasing.
To me, this is a net cognitive gain. It gives us room to think more clearly, creatively, and horizontally, to make connections across disciplines, to explore new angles, to develop ideas more authentically.
But this brings us to an important question: How do we know whether we’re using AI as an assistant—or as a replacement for our thinking?
As a rule of thumb:
- If AI is helping you express and refine what you have already thought, it’s assisting.
- If it’s generating ideas or content in your place, it’s replacing.
Here’s a table that helps clarify this distinction:
Final thoughts
As I often say, we are fortunate to be living through a moment of profound transformation—one where the rise of AI is reshaping how we think, work, and create. The generative AI revolution is not just a technological shift; it’s a cognitive and cultural one. But to truly benefit from its potential, we need to approach it with both wisdom and responsibility.
This means using AI not as a shortcut to bypass thought, but as a partner that challenges and refines it. When we treat AI as a collaborator, not a replacement, we open space for deeper inquiry, clearer expression, and more authentic insight. The challenge ahead is not just learning how to use AI, but how to stay intellectually present while doing so. That, as far as I am concerned, is the real test, and opportunity, of this new era.
References
Gerlich, M. (2025). AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking. Societies, 15(1), 6. https://doi.org/10.3390/soc15010006
Mollick, E. (2024). Co-intelligence: Living and working with AI. Little, Brown Spark.
Smith, G & Funk, J. (2024). When It Comes to Critical Thinking, AI Flunks the Test. The Chronicle of Higher Education. https://www.chronicle.com/article/when-it-comes-to-critical-thinking-ai-flunks-the-test
Vallor, S. (2024). The AI mirror: How to reclaim our humanity in the age of machine thinking [Kindle edition]. Oxford University Press
The post AI Assisting Thinking Vs Outsourcing Thinking to AI appeared first on Educators Technology.
Title: AI Assisting Thinking Vs Outsourcing Thinking to AI
URL: https://www.educatorstechnology.com/2025/04/ai-assisting-thinking-vs-outsourcing-thinking-to-ai.html
Source: Educational Technology
Source URL: https://www.educatorstechnology.com
Date: April 5, 2025 at 09:51PM
Feedly Board(s): Schule
Das Bundesinnenministerium befürchtet, dass Trump das EU-US Data Privacy Framework beenden könnte. Dann wäre Nutzung von US-Clouddiensten unmöglich. (
Wir besuchen im Podcast das Deutsche Museum und lassen uns Codes und Chiffren erklären. (
Die Visual Suite 2.0 von Canva ist fertig. Golem.de sprach mit dem Produktchef über neue KI-Funktionen und das Integrieren von Übernahmen. (
