Framework
Predictive versus generative AI
The question it answers
When a CEO is asked to bring AI into the business, how does a model that classifies and predicts differ from one that generates content, and which fits the job?
- Origin
- Feuerriegel et al., 2024; Vassilev et al., 2025
- Diagram type
- Comparison table
- Level of analysis
- Firm
Diagram
Predictive versus generative AI
| Criterion | Predictive AI (mostly discriminative) | Generative AI |
|---|---|---|
| What the model learns | Mostly just a decision boundary between classes | The distribution of the data, so it can produce new samples |
| Output | A label or value for new inputs: a discrete class (classification) or a continuous value (regression) | New content such as text, images, audio or code |
| Typical use | Data-driven decision support, for example detecting and classifying objects in images | Content creation and question answering, for example GPT-4, Dall-E 2, Copilot |
| Working with people | Delegation: a person hands a decision to the system | Co-creation: a person prompts and the output shapes the next prompt |
| Risk the sources name | Deliberately perturbed inputs can make the model misclassify, for example a stop sign read as a speed limit sign | Plausible but wrong output (hallucination), bias, copyright violation, energy use |
When to use it, and when not to
Use it when
In a study of 5,172 customer-support agents, a generative AI conversational assistant raised issues resolved per hour by 15 per cent on average, with the largest gains for less experienced workers (Brynjolfsson, Li and Raymond, 2025).
Pitfalls
Pitfall
The line is how the model learns, not the task: NIST notes that generative models can also be used for predictive tasks such as sentiment analysis.
Sources
References
Listed A–Z by author surname, Harvard style. A source with a DOI links to https://doi.org/…; one without links to its page or says plainly that it has no DOI.
- Brynjolfsson, E., Li, D. and Raymond, L.R. (2025) ‘Generative AI at work’, The Quarterly Journal of Economics, 140(2), pp. 889–942.doi:10.1093/qje/qjae044Cited in 1 article
- Feuerriegel, S. et al. (2024) ‘Generative AI’, Business & Information Systems Engineering, 66(1), pp. 111–126.doi:10.1007/s12599-023-00834-7Source page
- Vassilev, A. et al. (2025) Adversarial machine learning: a taxonomy and terminology of attacks and mitigations. Gaithersburg, MD: National Institute of Standards and Technology (NIST AI 100-2e2025).doi:10.6028/NIST.AI.100-2e2025Source page