Scribd
Upload
4 views1 page

Generative AI Explained

Generative AI creates new content based on prompts and learned patterns, producing outputs like text, images, and code. While it excels in tasks with clear context and human oversight, it can also generate errors and unsupported claims. Best practices suggest using generative AI as a collaborative tool rather than an unquestionable authority.

Uploaded by

hofeg36691
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
4 views1 page

Generative AI Explained

Generative AI creates new content based on prompts and learned patterns, producing outputs like text, images, and code. While it excels in tasks with clear context and human oversight, it can also generate errors and unsupported claims. Best practices suggest using generative AI as a collaborative tool rather than an unquestionable authority.

Uploaded by

hofeg36691
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Generative AI Explained

Generative AI Explained
How text, image, code, and media generation systems produce new outputs.

What generative AI does


Generative AI creates new content from prompts, examples, or instructions. It can draft text, summarize
documents, write code, design images, create synthetic voices, generate music, or transform data into different
formats. The output is new, but it is shaped by patterns learned during training.

How language models work


A large language model is trained to predict likely sequences of text. During training, it learns relationships
among words, concepts, formats, styles, and reasoning patterns. During use, it receives a prompt and
generates a continuation that is statistically likely and contextually relevant.

- Prompt: the user instruction or context given to the model.


- Context window: the amount of text the model can consider at once.
- Tokens: small pieces of text, such as words, word fragments, or punctuation.
- Temperature: a setting that influences how predictable or varied the output is.

Strengths
Generative AI is useful for first drafts, brainstorming, rewriting, translation, tutoring, code assistance, document
analysis, data transformation, and rapid prototyping. It is strongest when the task has clear context, a checkable
output, and a human reviewer who can judge quality.

Limitations
Generative AI can produce errors, fabricated details, biased language, or plausible but unsupported claims. It
may struggle with hidden assumptions, exact arithmetic, specialized facts, or requests that require current
information unless connected to reliable tools or sources.

Good use Risky use

Drafting an email that you will review Sending an unreviewed legal or medical answer

Summarizing a document you provide Assuming it knows the contents of an unseen


document

Brainstorming ideas Treating every idea as factually correct

Writing code with tests Deploying generated code without review

Best practice: treat generative AI as a collaborator that accelerates work, not as an unquestionable
authority.

Page 1

You might also like