In Between Social Science and Computer Science
Talking About Artificial Intelligence and Climate Change
For today I started writing an outline for a talk and I thought I would share it with you. I am not sure it is great or decent yet however I would love your opinion on it. It is not my intention to be overly moralising, simply to raise a few questions.
Outline of a Talk — Fifty Shades of AI
My talk is named Fifty Shades of AI.
Beyond making AI work.
When the algorithms run.
When the machine learning techniques are learning.
What then?
My name is Alex, I’m 28 and I am currently studying computer science and social science at the University of Oslo. I have had a personal project of writing about AI every day for 500 days and I call this 500 days of AI. I will be on day 100 next week. So I will try to touch briefly on three topics I believe to be of great importance in the present.
- Beyond operational AI
- Funding and self-policing of ethical conduct by large technology firms.
- As well as shifting the focus of the field of artificial intelligence from ease of use or extraction to addressing the climate crisis.
We ask ourselves the question of who an autonomous car should hit. When perhaps we should be asking which vehicles should be made autonomous?
It takes a lot of minerals to produce the graphics processing unit of a car or CPU and that means mining. How much mining activity will fleets of autonomous cars require to be maintained or to be operational?
If autonomous cars are running on the most efficient hardware how much energy does it take to constantly navigate the roads as opposed to the energy saved by more efficient driving?
Jeff Hammerbacher a prolific developer in his 20s said:
"The best minds of my generation are thinking about how to make people click ads. That sucks."
I do want to believe technology companies. I want to believe ‘don’t be evil’ as Google says; I want to believe in bringing the world closer together; I want to believe in being empowered to achieve more.
I want these companies to be ethical, to do the right thing in the right way for humanity in these times of a climate crisis. However we have seen examples before of self-policing ethics, corporate science. The tobacco industry funding scientific health research; the pharmaceutical industry regulating medical research; oil companies looking after their own oil spills; mining companies monitoring rivers or surrounding lands. The last thing we need is self-policing technology firms.
Facebook funds ethics institutes.
Google owned DeepMind builds evaluation tools for AI algorithm performance (bsuite, Google it). Is this somewhat problematic?
If at times natural science is so artificial and social science is not very socially responsible where does it leave us?
Saving the Amazon perhaps, not the company shipping products very quickly, no the Amazon forest nurturing life on earth.
Saving lives in hospitals with cognitive machine learning techniques while protecting privacy.
Thinking seriously and intently on data citizenship. Then making regulations and initiatives to ensure a higher degree of fairness.
To think that algorithms can show behavioural patterns and set out to change these and make them transparent.
Building new technology has been an interdisciplinary endeavour and it still is. Yet we need to go from thinking users to thinking people, citizens. Together we can work to bring both intelligence and understanding closer.
The fields of social science and computer science can work together and this collaboration will be important if we are to contribute in the time ahead. A time of crisis.
Technology cannot save the world, but it can assist in the process of making the world a bit better. You can make the world a bit better.
This is day 94 of #500daysofAI. My current focus for day 50–100 is on AI Safety. If you enjoy this please give me a response as I do want to improve my writing or discover new research, companies and projects.
