Aerospace Garage

Upskill / Business and Digital / Using AI tools safely at work in aviation

Using AI tools safely at work in aviation

What chat assistants, classifiers and predictive tools can and cannot do, checking their output against approved data, protecting data, human oversight and company policy. Part 2 of the AI in aviation series.

By Aerospace Garage Desk · Practitioner · 45 minutes · 24 narrated pages

What you will be able to do

  • ✓ Distinguish language models, classifiers and predictive models, and describe what each outputs and how each typically goes wrong.
  • ✓ Decide where AI tools help and where approved data, certifying staff and other authorised people must remain the authority.
  • ✓ Verify AI output against independent sources, including manual search answers and AI-drafted occurrence summaries.
  • ✓ Handle personal, confidential, proprietary and export-controlled data correctly when AI tools are used.
  • ✓ Recognise automation bias, using the official findings on Eastern Air Lines Flight 401, and design human checks that stay effective.
  • ✓ Write effective prompts, apply a company AI policy, record AI use and triage predictive maintenance alerts.

Course content

Session 1 · What today's AI tools are and are not

1.1

  1. Three families of AI tool · free preview
  2. How a chat assistant produces an answer
  3. Classifiers and predictive models
  4. Worked example: building Zellige Aero's AI tool register
Session 2 · Where AI helps and where it must not be the authority

1.2

  1. Where AI tools earn their place
  2. Approved data and the release decision
  3. Operations, quality, safety and airport work
  4. Worked example: checking an AI search of the maintenance manual
Session 3 · Hallucination and verification

1.3

  1. Why fluent answers can be false
  2. A verification routine you can use every time
  3. Summaries, translations and numbers
  4. Worked example: checking an AI-drafted occurrence summary
Session 4 · Data protection, confidential and export-controlled data

1.4

  1. Where your words go
  2. Personal data and safety reports
  3. Commercial and export-controlled technical data
  4. Worked example: what can go into which tool, under time pressure
Session 5 · Human oversight and automation bias

1.5

  1. Automation bias: why good people accept bad output
  2. What the regulators say about people and AI
  3. Keeping the human check real
  4. From the investigation file: Eastern Air Lines Flight 401
Session 6 · Using AI tools well: prompts, policy and records

1.6

  1. Prompting well
  2. Your company's AI policy
  3. Recording AI use and reporting its errors
  4. Worked example: triaging a predictive maintenance alert

Who this course is for

Upskill courses are for professional development and are not approved by any aviation authority. Learner Terms

Course content © Aerospace Garage Limited. All rights reserved.