Power, Politics and AI -
Bartoletti
Class Media technology & society
The age of AI has begun - Bill Gates
Positive view on the disruptive impact of AI
→ eg more productive health workers, help in triage, assist teachers in
grading & understanding student behavior
Computers were not as helpful for education as industry has hoped (cf give
every child a laptop)
→ new tech = always seen as ‘the solution’
→ now AI software will finally deliver on the promise
→ ignoring a lot of realities
Rise of AI will give people more free time
→ cf Bell in 1973
An Artificial Revolution - Ivana Bartoletti
What is AI?
→ difference between GPAI and …
→ we all talk about it, but we don’t really understand what it is
→ supposed to take over the world, solve all our problems, recreate human
life & make us all jobless?
⇒ definitions matter to understand it right
AI = machines perform a task that humans perform, bc we taught it how
to do it
→ machine that makes decisions, identifies anomalies & performs
specific tasks
→ eg very good at spotting anomalies in heart conditions (maybe even
better than humans), but it has to be taught
An Artificial Revolution: On Power, Politics and AI - Bartoletti 1
, Machine learning = labour-intensive activity
→ underpaid & unrecognized large number of workers
→ similar to traditional factory work (vs futuristic tech fantasies)
We are humanising robots + giving them some agency
→ huge leap into the unknown
Focus on robots, general intelligence etc = distraction
→ keep most people in ignorance
→ prevents us from asking who and what is AI for?
Repeating same mistakes as with the Internet
→ internet allows to explore new possibilities & be more connected
→ but its infrastructure has changed our privacy, our way of life &
control over our information
Problem now = what should we be worrying about?
First pages = explaining AI
→ we can train machines w/o bias (this is the idea)
The politics of data
Challenges basic idea that data can inform better decisions
→ concept = data mirrors society & reflects it honestly
⇒ sanctity of data (objective holy grail)
In reality = data is not neutral
→ collecting huge amounts brings many challenges
→ privacy & power dynamics are involved
Data-driven decisions are marketed as most neutral, objective, informative
& efficient
→ misleading bc someone trains an algorithm (aka inherent bias in the data
that is chosen)
An Artificial Revolution: On Power, Politics and AI - Bartoletti 2