No ratings yet.

When ChatGPT launched halfway through my first year of university, I was immediately curious. At the time, the tool was far from perfect. At first, the tool wasn’t particularly reliable, answers were often awkwardly phrased or simply wrong. Still, I did see potential. Over time, I watched it evolve into a far more capable assistant, able to support me in nearly every stage of student work, from brainstorming to proofreading.

My own use has fluctuated a lot. At times, I leaned on GenAI quite heavily, letting it help me rephrase sentences, brainstorm, or quickly explain course concepts. Other times, I tried to step back, aware that if I used the tool too much, too soon, I might shortcut my own learning. It can sometimes feel tempting just to take the instant answer and move on, but I try to remind myself that working through the material on my own is usually where the real learning takes place.

This tension between productivity and overreliance feels central to my experience. GenAI can save me enormous amounts of time and provide personalised support whenever I need it. Yet, I also worry about what researchers call “cognitive debt”: the idea that constant reliance on these tools can erode critical thinking and independence (Kosmyna et al., 2025). I sometimes catch myself leaning on the tool too quickly, and that awareness has pushed me to be more deliberate about when, why, and how I use it. Granted, I have to admit that this is the case 100% of the time. However, I do really try to use GenAI more consciously. For routine tasks like polishing language or checking clarity, it’s an excellent resource. But for deeper understanding, I make a point of working things out myself first. In that sense, it’s less about the technology itself and more about the habits I build around it. Used carelessly, GenAI can negatively impact the learning process. Used wisely, it can free up time and energy for the parts of study that really matter.

Reference

Nataliya Kosmyna, Eugene Hauptmann, Ye Tong Yuan, Jessica Situ, Xian-Hao Liao, Ashly Vivian Beresnitzky, Iris Braunstein, and Pattie Maes. Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task, June 2025. URL http://arxiv.org/abs/2506.08872. arXiv:2506.08872 [cs]

No ratings yet.

Last month, Meta began rolling out its AI assistant inside WhatsApp across Europe. The new feature appears as a blue circle at the bottom of users’ chat lists, offering capabilities like answering questions, generating images, and providing web search results. This follows the explosive growth of GenAI chat platforms, with Meta clearly aiming to capture some of that momentum. But the early reaction tells a different story, search results on the feature are dominated by tutorials on how to turn it off (Fraser, 2025).

This backlash highlights a core principle of digital strategy: business models succeed when they solve real customer problems, not when they adopt technology for its own sake. WhatsApp’s value proposition is simple and reliable communication. A permanent AI button now risks undermining those strengths. As McKinsey notes, true customer-centricity means redesigning processes around what customers need, not what new technologies make possible (Sankur et al., 2024).

And yet, dismissing the move outright may be short-sighted. In the long run, Meta could be positioning WhatsApp as a new kind of platform where an assistant helps users search, plan, and manage tasks without leaving the chat. If so, today’s complaints may be similar to the initial backlash against Instagram Stories. This rollout too faced temporary resistance but later on became central to the platform. To reach that goal, there are ways WhatsApp could still steer this rollout in the right direction. First of all, clear opt-out controls would give users a sense of agency. Equally important is rebuilding the trust around privacy. User fears about Meta AI scanning their messages aren’t new, previous privacy concerns already drove users to competitor Signal earlier this summer. Reaffirming how personal messages are never scanned could address those concerns. Finally, demonstrating useful, low-friction applications (like summarising long group chats or translating messages instantly) would help shift perceptions.

So is Meta forcing unwanted tech onto its users, or laying the groundwork for trusted new features? The answer will depend on whether Meta AI evolves into a genuinely helpful tool, or remains an unwanted button nobody asked for.

References

Fraser, G. (2025, 23 april). WhatsApp defends “optional” AI tool that cannot be turned off. https://www.bbc.com/news/articles/cd7vzw78gz9o

Sankur, A., Duncan, E., Cilento, G., & Fuchs, S. (2024, 10 december). True customer-centricity: An operating model for competitive advantage. Mckinsey.com. https://www.mckinsey.com/industries/industrials-and-electronics/our-insights/true-customer-centricity-an-operating-model-for-competitive-advantage