Talk:Issues in Interdisciplinarity 2018-19/Subjective and Objective Truth in AI

Minutes from first meeting: Monday 26th November
Present: Serene, Valentine, Jemima Meeting duration: 10am - 12pm

Discussed title of page and decided on Subjective and Objective Truth in AI - can change later if need be Brainstormed what we would write about:


 * Subjective truth in AI - any ethical decisions made by the AI itself, personality of the AI and any ethical decisions we make with regards to AI (playing God, biases coded into AI etc)
 * Objective truth in AI - machine learning algorithms, decision making, ending on bias, uses of AI (dangerous work environments etc)

For chapter:
 * define both subjective and objective truth
 * write about objective truth - AI is coded, scientific, seemingly objective, can learn continually
 * write about subjective truth - plays a big role, ethical decisions made by AI (self driving cars etc), touch on our own ethics if we have enough words
 * write about the contradictions between the two (can code be completely without bias?)

Possible subheadings: Jemimagbarnes (discuss • contribs) 10:45, 26 November 2018 (UTC)
 * Subjective vs. Objective Truth (definitions)
 * Objective truth within AI
 * Subjective truth within AI
 * Issues and contradictions (?)

Splitting up work and time plan

 * Subjective vs Objective - lifted from seminar sandbox. J to check for plagiarism
 * Objective Truth in AI - S to research and write
 * Subjective Truth in AI - V to research and write
 * Issues and Contradictions - J to research and write

ALL - when written go through and check for plagiarism, check references are all correct and linked properly, check for coherence and cut down words if required.

Timescales:

Jemimagbarnes (discuss • contribs) 11:36, 26 November 2018 (UTC)
 * All sections written by 3rd December
 * Cut down and proof read by 10th December

Edit
Altered the definition of objective truth to make it more succinct and clear

Finished section on Objective Truth in AI

Altered the definition of subjective truth to make it more succinct and clear

Cancelled out content due to word count limit

Serenechan3 (discuss • contribs) 12:06, 4 December 2018 (UTC)

Suggestions
Structure Jemimagbarnes (discuss • contribs) 15:31, 4 December 2018 (UTC)
 * Definitions and where they are from. eg. The accepted definitions of subjective and objective truth in philosophy are being used here
 * Objective truth - AI is objective as uses raw data sets, learns from data sets that don't work etc and can repeat things over and over in the same conditions (unlike humans). In this way it can be considered objective. Use example from Accenture.
 * Subjective truth - AI might not be as objective as we think. eg. It has been suggested that... Use more of a passive voice.
 * Issues - need to make absolutely clear that AI would benefit from an interdisciplinary approach

Rewriting stuff
Thread of the chapter could read in this kind of way:

- AI is very objective because of all the maths and data and logic.

- AI is not that objective because the data is biased and its coded by biased humans + it has to make ethical decisions

- We need to look at AI from lots of different angles or else we're going to end up with a completely unethical 'black box' that is just processing our own biases and worst bits from humanity

Link to rewritten chapter: https://liveuclac-my.sharepoint.com/:w:/g/personal/uclqjgb_ucl_ac_uk/EWEa4yoy3ddLktMLiM8osKkB4mo29nQzDo4oL8Z5Z5iPXQ?e=Ys5KvI

Jemimagbarnes (discuss • contribs) 12:33, 5 December 2018 (UTC)

Minutes from second meeting: Friday 7th December 2018
Present: Serene, Valentine, Jemima Meeting duration: 2:30pm - 3:30pm

Tasks:

Finalize chapter and put it into Wikibook

Complete references in Vancouver style