Chapter 10 — Engineering ethics and safety
Lesson map
Lesson 10 published. Map, example, EX028.
flowchart TD
N0["Learning outcomes"]
N1["10.1 Duties"]
N2["10.2 Privacy"]
N3["10.3 Authorisation"]
N4["10.4 Safety"]
N5["10.5 Raising concerns"]
N6["10.6 Honesty"]
N7["Professional codes"]
N0 --> N1 --> N2 --> N3 --> N4 --> N5 --> N6 --> N7
N7 --> EX["Worked example"]
EX --> Q["Exercise EX028"]
Logical drawing
flowchart LR
N0["Learning outcomes"]
N1["10.1 Duties"]
N2["10.2 Privacy"]
N3["10.3 Authorisation"]
N4["10.4 Safety"]
N5["10.5 Raising concerns"]
N6["10.6 Honesty"]
N7["Professional codes"]
N0 --> N1 --> N2 --> N3 --> N4 --> N5 --> N6 --> N7 --> Q["EX028"]
[Learning outcomes]
|
v
[10.1 Duties]
|
v
[10.2 Privacy]
|
v
[10.3 Authorisation]
|
v
[10.4 Safety]
|
v
[10.5 Raising concerns]
|
v
[10.6 Honesty]
|
v
[Professional codes]
|
v
[Worked example] --> [Exercise EX028]
Learning outcomes
You will (1) apply engineering ethics to cloud work; (2) consider safety and privacy of households; (3) act within authorisation; (4) raise concerns responsibly.
10.1 Duties
Engineers owe duties to the public, clients and the profession. For meterhub, the public includes 80,000 households whose readings reveal when they are home.
The public's interest is not abstract here: a careless design could expose patterns of daily life for tens of thousands of families. Clients expect competent work within agreed scope. The profession expects honesty and continued learning. When these duties conflict, the public interest comes first. Elena writes these duties into her project plan so they guide daily choices, not only difficult moments.
10.2 Privacy
Meter readings at fine intervals can reveal occupancy and habits. Elena's project uses synthetic readings; real data stays in production under Solent's controls.
Synthetic readings let Elena build and test without touching personal data, which removes a whole class of risk from the project. Real data stays where Solent's access controls, logging and retention rules apply. If a test truly needs real data, it goes through Solent's approval process first. Elena's synthetic generator mimics evening peaks and holiday dips without copying any real household's pattern.
10.3 Authorisation
Elena's access is limited to read-only exports and a test cluster. She does not test production systems without written approval from Rosa.
Read-only exports and a test cluster give Elena enough to do her work without risking production. Written approval for any production test makes the decision explicit and traceable. If something goes wrong during an approved test, the record shows it was sanctioned and by whom. She keeps copies of each approval with the test records. Approvals state the date, scope and stop conditions, and expire after the test window ends.
10.4 Safety
Changes to network policies could stop ingestion, which affects billing and grid planning. Safety means staged rollout, monitoring and rollback plans.
Ingestion stopping means readings are lost or delayed, which affects bills and the grid operator's planning. Staged rollout limits the impact to a few services first; monitoring shows problems quickly; rollback restores service if needed. Elena writes these steps into every change plan. She also checks each change against the evening peak and avoids rolling out during it, when readings are most valuable.
10.5 Raising concerns
If Elena found that readings were exposed, she would report to Rosa and Solent's data protection lead immediately rather than investigate alone.
Investigating alone could destroy evidence, alert an attacker or breach rules on handling personal data. Reporting immediately lets people with authority and expertise take over. The data protection lead also judges whether a regulator must be notified, which has its own deadlines. Solent's incident procedure lists who to call and how quickly, and Elena keeps a copy handy.
10.6 Honesty
Reports state uncertainty and do not overstate the effect of controls.
Saying a control 'prevents' an attack when it only makes one harder misleads decision-makers. Elena uses measured results where possible and states what was not tested. Honest reports build trust and help Solent decide where further work is needed. When results are mixed, the report gives both sides and the evidence for each. Elena also lists known limitations at the start of each results chapter, not just at the end, so readers meet them before forming conclusions. Rosa checks reports for overclaiming before they reach managers.
Professional codes
Elena reads her engineering institution's code of conduct and notes three duties relevant to her project: protect public safety and privacy, work only within competence, and report concerns. Her project plan has a short ethics section that names these duties and how she meets them: synthetic data, read-only access, staged changes and clear escalation contacts. Dr Wagner signs the section at the start of the project.
Ethics in design
Ethics also shapes design choices: Elena prefers controls that fail safely and are visible to teams, rather than hidden controls that could surprise people and cause outages.
Worked example
During the project, Elena notices a test dataset in the shared cluster that looks like real readings with addresses. She reports it; Solent confirms it was a copy made against policy, deletes it and reminds teams of the synthetic-data rule.
Synthetic case. Events are synthetic.
Errors and limits
Good intentions do not replace authorisation.
Exercises
- EX028 — Public duty. Who is the public for meterhub? Rubric: 3 = 80,000 households; 2 = partial; 1 = vague; 0 = none.
- EX029 — Privacy risk. What can fine-grained readings reveal? Rubric: 3 = occupancy and habits; 2 = partial; 1 = vague; 0 = none.
- EX030 — Real data found. What did Elena do on finding real-looking data? Rubric: 3 = reported to Rosa; 2 = partial; 1 = vague; 0 = none.
Synthesis
Ethics guides how engineers handle data and changes.
Continue studying — chapter 10 of 50
← Chapter 09 (draft) · Course chapter index · Chapter 11 (draft) →