Term

Generative AI

Generative AI is the class of AI systems whose output is new content rather than a label or a score. A predictive system looks at an input and answers a closed question (is this email spam, will this customer churn). A generative system looks at an input and produces an artifact (a paragraph, an image, a snippet of code, a synthesized voice). The substrate underneath almost all of it is the same: large foundation models, mostly transformer-based, trained on broad unlabeled corpora and then adapted to specific tasks.
Reviewed by 7wData
Mentioned Large Language Model (LLM) Hallucination (in LLMs) Prompt Injection

Why it matters

The wave of enterprise AI adoption from 2023 through 2026 is overwhelmingly GenAI-shaped. When a board asks about “AI strategy” they mean GenAI. When a vendor pitches an “AI feature” they mean GenAI. Most of the active “AI governance” conversation is really GenAI governance, because the older predictive ML stack already had its own quieter controls. The output modalities are text via LLMs, image, audio, video, and code.

Where you’ll encounter it

Three contexts. A vendor pitches “GenAI for X” and the honest reading is almost always an LLM wrapper with a system prompt and a retrieval step, not a custom-trained model. A procurement question asks “do you use GenAI in your product” and the honest answer is more nuanced than yes or no, because calling a GenAI API at runtime and shipping GenAI output to the end user are different exposures. A regulator asks for documentation on GenAI-specific risks and expects to see how you handle hallucination, prompt injection, and training-data provenance.

Part of the 7wData AI Glossary. Tracking how concepts like this move in the expert conversation: daily signals at ins7ghts.com.