Ai and development

Question 1

‘AI-as-a-process’ is a sociotechnical device that
automates last-mile tasks

Answer 1

(Ludec et al, 2023) e.g ai cars, ai surveillance, self checkout

Question 2

Technological ‘fixes’

Answer 2

Drengson (1984) We should use technology to improve our daily lives and solve problems

Question 3

Techno-optimism

Answer 3

Danaher (2022) Faith in technology

Question 4

Techno-solutionism entails a faith in technology, but
also a tendency to fundamentally change how we
perceive and analyse social phenomena

Answer 4

(Morozov, 2013)

Question 5

How ai could help climate change?

Answer 5

WEF (2024) Knows where icebergs are melting and how fast,
Map deforestation
Help communities in Africa facing cc risks e.g predict weather patterns and improve access to clean energy

Question 6

AI for Good

Answer 6

UN (2023) Help meet SDG’s
e.g Disaster prevention
Carbon neutrality
Tracking pollution
Food production
‘Ai can optimise grids’

Question 7

‘AI can be a transformative tool in our fight against climate change’

Answer 7

Witherspoon from Google AI lab

Question 8

Supply chain of Ai

Answer 8

Valdivia (2024)

Question 9

GPUs are made of large variety of
minerals but silicon and copper are the
main ones – amongst gold, tantalum,
palladium, boron or tungsten

Answer 9

Euromines (2022)

Question 10

Case Study Queretaro Mexico

Answer 10

Valdivia (2023)
* Emerging as a hub of data centres
* 10 data centres with 18 additional
projects

Question 11

Case study factors

Answer 11

• Factors:
  ➢ Industrial legacy:
• US automotive factories to aerospace
  industry
  ➢ Geographical location:
• 2-hours from Mexico city
• Seas cables facilitate low latency for
  data centre connections
  ➢ Local administration: willing

Question 12

E-Waste

Answer 12

Technology Review

• 100% of Ethernet cables, networking
  components, servers and IT storage could end
  up in e-waste treatment plans (Whitehead et
  al., 2015)
• GPUs have a lifespan of 3 to 5 years
• Batteries or pipes have a refresh rate of 20
  years - landfills

Question 13

Labour behind AI production

Answer 13

Most work done in Global South

Poor conditions in Kenya
-Complaints of poor mental health support
Venezuela issues of private images being online

Many paid less than 2 euros per/h

FEEDING THE MACHINE- Cant, Muldoon, & Graham (2024)

Question 14

Environmental impacts

Answer 14

Muldoon and Wu (2023)

• A single natural language processing model produced 660,000 pounds of emissions, amounting to as
  much as five cars over their lifetime (Strubell et al. ,2019)
• At Google, machine learning accounts for 15% of the company’s total energy consumption
  (Patterson et al., 2022).
• Under-water cables and infrastructures

Question 15

Ai coloniality

Answer 15

After Muldoon and Wu, 2023 (Page – 10)
* The environmental costs of this technology are not
distributed equally, with countries vulnerable to climate
change related catastrophes most at risk (Bender et al.,
2021; Westra & Lawson, 2001).
* Technological growth also increases the amount of e-waste
the world produces, which is another cost
disproportionately shouldered by the majority world. Ewaste increased to 6.8 kg per capita in 2021, with long-term
estimates predicting over 120 million metric tons of e-waste
per year by 2050 (Dauvergne, 2022).