AI in Product Design - Morgenbooster 11/09/2024

There is no doubt that AI capability will be a key feature of many products in the future. But does this impact the way we design products? And how can we avoid designing for AI's sake, but design with (the right) intent?

1. AI in
Product
Design
Morgenbooster 11.09.2024

3. Linnea Forss
Senior User Experience Designer
Lisbeth Torp Christensen
Head of Digital Products

4. ​
01   What is AI good at
​
02  AI in the design process
​
03  Responsible AI
Agenda

5. As tools
A helper for us to get better, more creative
and faster in our work
As product
AI Incorporated into our products to create
a better experience and better flows
AI

6. ​
"Artificial intelligence is the science and engineering of making
​
intelligent machines"
​
John McCarthy, Professor at Stanford, 1955

7. What is AI good at?

8. ​
5 types of AI
​​
Predict
​​
Personalize
​​
Recognize
​​
Generate
​​
Uncover
structure
​

9. ​
Example
​
Uncover structure

10. ​
Example
​
Predict

11. ​
Example
​
Personalize

12. ​
Example
​
Recognize

13. ​
Example
​
Generate

14. AI in the design process

15. Think
What’s the problem
really?
Create
Explore, prototype,
test, learn.
Build
Develop to evolve
User
Tech

16. Users first and
technology,
also first
​

17. Think

18. ​
Starting with user
​
pain points
Look for pain points that are
.. Timeconsuming and sequenced
.. Repetitive
.. Labor-intensive
.. Overwhelming
.. Generic or impersonal

19. ​
Starting with user
​
pain points
Look for pain points that are
.. Timeconsuming and sequenced
.. Repetitive
.. Labor-intensive
.. Overwhelming
.. Generic or impersonal

20. ​
From pain points to
opportunities
Use ideation cards on top of current journey to
identify where processes could be optimized and
what AI capabilities might be relevant.
​

21. ​
From pain points to
opportunities
Use ideation cards on top of current journey to
identify where processes could be optimized and
what AI capabilities might be relevant.
​
AI knows the
customer
G A I N S I N S I G H T S F R O M B I G D ATA A N D
UNDERSTANDS THE PAST AND PRESENT
12
AI highly
personalizes
products and
services
M A K E S D E C I S I O N S A N D
RECOMMENDATIONS
22
AI plans next
best actions
M A K E S D E C I S I O N S A N D
RECOMMENDATIONS
20

22. ​
Starting
​
with data
Start mapping data you have and
explore what data could be useful
to get.

23. ​
Lightning demos
in ideation
Be inspired from other AI solutions:
Internal feature catalog

25. Create
Test
Design
Evaluate

26. Ideate
Concepts

27. Evaluate
Data quality
​
Availability
Is it possible to retrieve?
Reliability
Is it complete? Representative?
Frequency
How often is data updated?
Cost
How expensive is it to collect and
maintain?
Precision
How close are the data values to
the expected values?
Age + location +
Other people’s
recommendations

28. Design the
behaviour of
the AI
​

29. Design the
behaviour of
the AI
​

30. Test the
‘old’ methods
Source: NN Group

31. Test the
POC

32. Build
Evaluate
Evaluate
Evaluate

33. ​
Design challenges
​​
User control
​​
Unforeseen
consequences
​​
Planetary
costs
​​
Trust
​​
Manipulation
& bias
​

34. Evaluating concepts
Viability Feasibility
Desirability
Concept

35. Make room for
responsibility
Responsibility
Viability Feasibility
Desirability
Concept

36. Responsibility
people

37. ​
People principles
​​
Make the technology
understandable
​
Give users the control
​
Avoid manipulation and bias
! ✊ #

38. Make the
technology
understandable
✅⛔

39. Give users
the control
✅⛔

40. Avoid
manipulation
and bias
✅⛔

41. ​
The Digital Ethics
Compass
​

42. Responsibility
planet

43. ​
Planet principles
​​
Be resource efficient
&
​
Use resources carefully and improve AI systems to
save energy and work more efficiently.

44. ​
Planet principles
​​
Think small
​
Choose your partners
with care
​
Keep monitoring
' ( )

45. Think small
​

46. Choose your
partners with care
​

48. Keep monitoring

50. ​
Responsible AI principles
​​
Think small
​
Choose your partners
with care
​
Keep monitoring
' ( )
​
Make the technology
understandable
​
Give users the control
​
Avoid manipulation and bias
! ✊ #

51. ​
Two future perspectives
​

​

52. ​
"Your future is whatever you
make it. So make it a good one,
both of you."
​

53. ​
"Your future is whatever you
make it. So make it a good one,
​​
all of you."

54. ​
Resources
​
General design ressources for AI:
AI meets design toolkit
https://aixdesign.co/posts/ai-meets-design-toolkit
Digital Ethics Compass
https://ddc.dk/tools/toolkit-the-digital-ethics-compass
33a cards for ideation
https://www.33a.ai
Google People + AI guidebook with exercises
and advice
https://pair.withgoogle.com/guidebook
​
Responsible AI ressources:
Responsible AI one-pagers (under construction)
The digital ethics compass:
https://ddc.dk/tools/det-digitale-etikkompas-
vaerktojskasse/
Find green hosting providers:
https://app.greenweb.org/directory/
Monitoring tools tools:
https://codecarbon.io/
https://www.green-algorithms.org/
https://mlco2.github.io/impact/

55. Course:
Designing
responsible AI
products
07.10.2024
[email protected]

56. Thank you