liveBook · Manning

Ethics

As explained in Build an Advanced RAG Application (From Scratch)

A moral or societal issue arising from the potential consequences or use of an artificial intelligence system.

Ethics concerns the moral and societal consequences of deploying language models. It includes privacy, consent, misuse, and the risk that generated content causes harm or spreads misinformation. section 1.3 section 5.5

Responsible use requires attention to related failure modes:

  • Bias can reproduce stereotypes and produce unfair or discriminatory outputs. section 1.3 section 5.5
  • Hallucinations can present fabricated information as plausible fact, creating risks in critical applications. section 1.3 section 5.5
  • Prompt hacking can manipulate a model into producing inappropriate, biased, or misleading content. section 5.5
  • Limited interpretability can weaken trust and transparency when users cannot understand generated responses or decisions. section 1.3

Prompt manipulation and misinformation

A health-advice system may be manipulated by a prompt asking it to justify skipping vaccinations. If it follows the manipulated framing, its response could spread misinformation despite an intended focus on accurate advice. section 5.5

Ethical deployment therefore depends on monitoring, regulation, validation, and fact-checking. These controls help reduce the risks of harmful, misleading, biased, or unreliable outputs. section 1.3 section 5.5

How Build an Advanced RAG Application (From Scratch) covers it

How deep it goes
Identifies major ethical risks and calls for monitoring and regulation, but stops short of specifying an operational ethics framework.
Examples
Uses vaccination misinformation from prompt manipulation and fabricated claims about a first person on Mars as concrete risk examples.
What you can do after
Evaluate LLM deployments for ethical risks and apply monitoring, regulation, validation, and fact-checking as mitigation practices.
How it connects
Connects ethics with bias, privacy, consent, misuse, hallucinations, prompt hacking, interpretability, validation, and fact-checking.

Also explained in

Compare how 4 books explain ethics →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build an Advanced RAG Application (From Scratch). The section links point to the parts of the book it draws on.