Designing technology ethically means considering its impact beyond just whether it functions correctly — questions like data privacy, algorithmic bias (where a system produces systematically unfair outcomes for certain groups), accessibility for people with disabilities, and the broader social consequences of widespread adoption. Sustainability in technology design considers a product's full lifecycle environmental impact: the resources and energy used in manufacturing, the energy consumed during use, and what happens to the device at end-of-life (e-waste) — with 'planned obsolescence' (designing products to become outdated or fail sooner than technically necessary) representing a significant, genuinely debated ethical and environmental issue in the industry.
Example
A facial recognition system trained mostly on images of one demographic group might perform significantly less accurately for people outside that group — a real example of algorithmic bias with serious consequences if the system is used for something like security or law enforcement, illustrating why considering diverse, representative training data is an ethical design responsibility, not just a technical afterthought.
Key terms
Algorithmic bias:
When a system produces systematically unfair outcomes for certain groups.
Planned obsolescence:
Designing products to become outdated or fail sooner than technically necessary.
Questions
1. Ethical technology design considers:
Impact beyond just whether the technology functions correctly
Only whether the code compiles without errors
Nothing beyond basic functionality
Only a product's visual design
2. Algorithmic bias occurs when a system:
Produces systematically unfair outcomes for certain groups
Always treats every single group with perfect, identical fairness
Has no possible connection to fairness or outcomes
Only ever affects hardware, never software
3. Sustainability in technology design considers:
A product's full lifecycle environmental impact
Only how a product looks when new
Nothing related to environmental impact
Only the cost of a product to the consumer
4. Planned obsolescence means:
Designing products to become outdated or fail sooner than necessary
Designing products to last as long as technically possible
A concept with no connection to product design
Only relevant to food products, not technology
5. E-waste refers to:
Discarded electronic devices and their environmental impact
A type of encrypted data file
Something unrelated to technology at all
Only paper waste from printing
6. A facial recognition system performing less accurately for certain groups is an example of:
Algorithmic bias
Perfect, unbiased system design
Something unrelated to fairness or accuracy
A hardware manufacturing defect only
7. Accessibility in technology design means considering:
Usability for people with disabilities
Only the needs of the most technically skilled users
Nothing related to who can actually use a product
Only a product's price
8. Why might training an AI system on data that isn't representative of the full population it will be used on lead to biased outcomes?
A system learns patterns from its training data, so if that data over-represents some groups and under-represents others, its outputs are more likely to be less accurate or fair for the under-represented groups
The data used to train an AI system has no genuine bearing on whether its resulting outputs are fair or biased
A system will always automatically produce fair and unbiased outcomes regardless of what data it was trained on
Representative training data has no connection to how accurately a system performs across different groups
9. Why might a technology company face genuine ethical criticism if it designs a product with planned obsolescence, even if doing so is legal?
Deliberately shortening a product's useful life can be seen as prioritising profit over consumer interests and environmental impact, raising ethical (not just legal) concerns
Legality and ethics are always exactly the same thing, so anything legal can never face any genuine ethical criticism
Planned obsolescence never actually raises any ethical concerns, regardless of its environmental or consumer impact
Consumer interests and environmental impact have no genuine connection to the ethics of product design choices
10. Why might considering a product's FULL lifecycle (manufacturing, use and disposal) give a more complete picture of its environmental impact than only considering its energy use while operating?
Manufacturing and disposal can involve significant resource extraction, energy use and waste that operational energy use alone doesn't capture
A product's environmental impact is always determined entirely by its energy use while operating, with manufacturing and disposal being irrelevant
Manufacturing and disposal processes never actually have any meaningful environmental impact worth considering
Considering the full lifecycle of a product provides no additional insight beyond simply measuring its operational energy use
11. Why might designing accessible technology (usable by people with disabilities) benefit a much broader range of users than just the specific group it was designed for?
Accessibility features (like voice control or larger text options) often improve usability for many other users too, not just the people they were originally designed for
Accessibility features designed for people with disabilities never actually benefit any other type of user in any way
Designing for accessibility always makes a product significantly worse or less usable for people without disabilities
Accessibility considerations in technology design have no genuine connection to the broader usability of a product
12. Why might a smartphone manufacturer choosing to use easily replaceable, standard-sized components (rather than proprietary, glued-in parts) be considered a more sustainable design choice?
Standard, replaceable components make repair more accessible and affordable, extending a device's useful life and reducing how quickly it becomes e-waste
The specific way internal components are designed and assembled has no genuine bearing on a device's environmental impact or lifespan
Proprietary, glued-in components always extend a device's useful life more effectively than standard, replaceable ones
Component design choices are always completely unrelated to how easily or affordably a device can eventually be repaired
13. Why might a company's decision about how long to continue providing software updates for an older device be considered both an ethical and environmental design choice, not just a business decision?
Ending support prematurely can leave a functional device insecure or unusable, effectively forcing earlier disposal and contributing to e-waste, beyond just its impact on the company's bottom line
Software update support timelines are always a purely internal business decision with no genuine ethical or environmental dimension
Ending support for an older device never actually has any real effect on whether a user eventually replaces it
The length of time a company supports a device with updates has no genuine connection to environmental sustainability
14. Why might a company be motivated to address algorithmic bias in its products even beyond genuine ethical concern, from a purely business perspective?
Biased outcomes can create legal liability, reputational damage, and lost trust or business from affected users, giving companies a practical business incentive alongside ethical ones
There is never any practical business reason for a company to address algorithmic bias in its products
Legal liability and reputational damage have no genuine connection to a company's handling of algorithmic bias
Ethical and practical business motivations for addressing bias are always completely unrelated to each other
15. Why might addressing e-waste require cooperation between manufacturers, consumers and governments, rather than any single group alone being able to solve it?
Manufacturers influence product design and repairability, consumers influence purchasing and disposal habits, and governments can set regulations and recycling infrastructure — each plays a distinct, necessary role
E-waste is a problem that any single one of these groups could fully and completely solve entirely on their own
Manufacturers, consumers and governments each have no genuine role to play in addressing the problem of e-waste
Cooperation between different groups provides no additional benefit over relying on just one group to address e-waste alone
16. Why might "right to repair" movements (advocating for consumers' ability to repair their own devices) be considered directly connected to both the ethics and sustainability of technology design?
Restricting repairability can be seen as prioritising a manufacturer's commercial interests over consumer autonomy, while also increasing e-waste by making it harder to extend a device's useful life
The right to repair movement has no genuine connection to either the ethics or the sustainability of technology design
Restricting a consumer's ability to repair their own devices always has an identical, purely neutral effect on sustainability
Consumer autonomy and environmental impact are always completely unrelated considerations in discussions about product repairability
17. Why might addressing algorithmic bias require more than simply adding more data, if the underlying data itself reflects existing societal biases and inequalities?
If historical or societal data itself contains biased patterns, an algorithm trained on it can learn and perpetuate those biases regardless of how much of that same biased data is added
Adding more data to a system always automatically and completely eliminates any risk of algorithmic bias regardless of the data's underlying quality
Societal biases and inequalities never actually get reflected or captured within data used to train technological systems
The quality and fairness of underlying training data has no bearing on whether adding more of it can resolve algorithmic bias
18. Why might independent, external audits of an algorithm's outcomes be considered a more reliable way to identify bias than a company simply reviewing its own system internally?
An external, independent perspective is less likely to be influenced by internal assumptions or incentives that might otherwise make certain biases harder for the original developers to notice
A company reviewing its own system internally always identifies exactly the same biases as an independent external audit would
External audits of an algorithm's outcomes never actually provide any additional insight beyond an internal review
The independence of who is conducting a bias audit has no genuine bearing on how reliable its findings are likely to be
19. Why might a truly sustainable approach to technology design need to consider not just an individual product's footprint, but also broader systemic effects, like how easy a product is to repair or how long a company commits to supporting it with updates?
Repairability and long-term support commitments directly affect how long a product remains useful, which has a much larger cumulative effect on total resource use and e-waste than the footprint of manufacturing a single unit alone
A product's individual manufacturing footprint is always the only environmentally relevant factor, with repairability and support commitments being irrelevant
How long a company commits to supporting a product with updates has no genuine connection to its overall environmental impact
Considering broader systemic factors beyond a single product's footprint provides no additional insight into genuine technology sustainability
20. Why might involving people from diverse backgrounds in the design and testing process of a new technology help reduce the risk of unintentionally excluding or disadvantaging certain groups?
A design team with limited diversity may unintentionally overlook needs, use cases or potential harms that aren't part of their own direct experience, which a broader range of perspectives is more likely to catch
The diversity of a design team has no genuine bearing on whether a technology unintentionally excludes or disadvantages any particular group
Involving diverse perspectives in design and testing never actually helps identify any potential issues before release
A design team's own direct experience is always sufficient to fully anticipate the needs of every possible user group
21. Understanding ethics and sustainability in technology design mainly helps you to:
Evaluate the broader social and environmental responsibilities involved in designing and building technology
Assume ethical and environmental considerations are irrelevant to whether technology functions correctly
Ignore how training data can influence whether a system produces biased outcomes
Treat a product's environmental impact as limited only to its energy use while operating
Answer key (parent copy)
1. Impact beyond just whether the technology functions correctly
2. Produces systematically unfair outcomes for certain groups
3. A product's full lifecycle environmental impact
4. Designing products to become outdated or fail sooner than necessary
5. Discarded electronic devices and their environmental impact
6. Algorithmic bias
7. Usability for people with disabilities
8. A system learns patterns from its training data, so if that data over-represents some groups and under-represents others, its outputs are more likely to be less accurate or fair for the under-represented groups
9. Deliberately shortening a product's useful life can be seen as prioritising profit over consumer interests and environmental impact, raising ethical (not just legal) concerns
10. Manufacturing and disposal can involve significant resource extraction, energy use and waste that operational energy use alone doesn't capture
11. Accessibility features (like voice control or larger text options) often improve usability for many other users too, not just the people they were originally designed for
12. Standard, replaceable components make repair more accessible and affordable, extending a device's useful life and reducing how quickly it becomes e-waste
13. Ending support prematurely can leave a functional device insecure or unusable, effectively forcing earlier disposal and contributing to e-waste, beyond just its impact on the company's bottom line
14. Biased outcomes can create legal liability, reputational damage, and lost trust or business from affected users, giving companies a practical business incentive alongside ethical ones
15. Manufacturers influence product design and repairability, consumers influence purchasing and disposal habits, and governments can set regulations and recycling infrastructure — each plays a distinct, necessary role
16. Restricting repairability can be seen as prioritising a manufacturer's commercial interests over consumer autonomy, while also increasing e-waste by making it harder to extend a device's useful life
17. If historical or societal data itself contains biased patterns, an algorithm trained on it can learn and perpetuate those biases regardless of how much of that same biased data is added
18. An external, independent perspective is less likely to be influenced by internal assumptions or incentives that might otherwise make certain biases harder for the original developers to notice
19. Repairability and long-term support commitments directly affect how long a product remains useful, which has a much larger cumulative effect on total resource use and e-waste than the footprint of manufacturing a single unit alone
20. A design team with limited diversity may unintentionally overlook needs, use cases or potential harms that aren't part of their own direct experience, which a broader range of perspectives is more likely to catch
21. Evaluate the broader social and environmental responsibilities involved in designing and building technology