How does ChatGPT actually work?

AI has become integral for many, from chatting with ChatGPT to coding with DeepSeek. Behind the interfaces, complex engines are often misunderstood.

How does ChatGPT work?

ChatGPT means Chat Generative Pre-Trained Transformer. 'Generative' means it is able to “produce” content; 'Pre-trained' indicates it learned from large datasets before use; 'Transformer' is its neural network architecture that understands sentence context by analyzing all words simultaneously.

The AI Spectrum

AI types vary: Generative AI creates text, images, or music; Large Language Models (LLMs) like GPT-4 process and generate human language; Predictive AI forecasts outcomes using historical data.

Pros, cons, and what lies ahead.

AI offers benefits like increased productivity and rapid discoveries, but experts warn about 'trustworthy AI.' Since models often act as 'black boxes,' they can reflect biases or 'hallucinate' facts. This could lead to misinformation if users do not verify the information received while searching on generative AI.

Recent reports reveal a 'crisis of credibility' as deepfakes blur reality, prompting new laws such as Singapore's Elections (Integrity of Online Advertising) Act. As AI advances, developing literacy to verify AI-generated info is crucial. Keep yourself informed online and remember that AI can be a bane or a boon - it all depends on how you utilise it.

Writer: Sherlyn

Published: 27/04/2026

References

Islam, M. B. E., Haseeb, M., Batool, H., Ahtasham, N., & Muhammad, Z. (2024). AI Threats to Politics, Elections, and Democracy: A Blockchain-Based Deepfake Authenticity Verification Framework. Blockchains, 2(4), 458-481. https://doi.org/10.3390/blockchains2040020 

Kalota, F. (2024). A Primer on Generative Artificial Intelligence. Education Sciences, 14(2), 172. https://doi.org/10.3390/educsci14020172 

Laizure, S. C. (2024). Caution: ChatGPT Doesn’t Know What You Are Asking and Doesn’t Know What It Is Saying. The Journal of Pediatric Pharmacology and Therapeutics, 29(5), 558-560. https://doi.org/10.5863/1551-6776-29.5.558 

Lee, J., Lee, J., & Yoo, J. (2025). The role of large language models in the peer-review process: opportunities and challenges for medical journal reviewers and editors. Journal of Educational Evaluation for Health Professions, 22, 4. https://doi.org/10.3352/jeehp.2025.22.4 

Sufi, F. (2024). Generative Pre-Trained Transformer (GPT) in Research: A Systematic Review on Data Augmentation. Information, 15(2), 99. https://doi.org/10.3390/info15020099