Vox article on bias in algorithms Flashcards
What are “machine learning” tools ?
(machine learning is a type of artificial intelligence), it’s better to think about the idea of “training.” This involves exposing a computer to a bunch of data — any kind of data — and then that computer learns to make judgments, or predictions, about the information it processes based on the patterns it notices
What is a big issue with machine learning?
these systems are often applied to situations that have much more serious consequences than determining what a book is, and in scenarios where there isn’t necessarily an “objective” answer.
explains that we can think about algorithmic bias in two primary ways:
accuracy and impact
Accuracy:
An AI can have different accuracy rates for different demographic groups.
Impact:
Similarly, an algorithm can make vastly different decisions when applied to different populations
(impact diff communities differently)
How can AI be biased when unleased on the internet?
by teaching an artificial intelligence to crawl through the internet — and just reading what humans have already written — the system would produce prejudices against black people and women
What is Lily Hu’s opinion on the fundamental problem of machine learning?
“You don’t have any guarantees because your algorithm performs ‘fairly’ on your old dataset,” Hu told Recode. “That’s just a fundamental problem of machine learning. Machine learning works on old data [and] on training data. And it doesn’t work on new data, because we haven’t collected that data yet.”
(works on old biased data sets)
What is an intersectional critique on monitoring AI?
A tool may deemed fairly accurate on white women, for instance, but that doesn’t necessarily mean it works with black women.
What is a potential issue with the designers of AI?
AI tends to be dominated by men. And the “high tech” sector, more broadly, tends to overrepresent white people and underrepresent black and Latinx people
- we need to think about who gets a seat at the table when these systems are proposed
Explain the issue of: Which questions we can even us AI for?
Ex. Ai wont ultimately address the ethical question of whether someone deserves parole
When you then insert an AI system, an algorithmic system, [or] a computer, that doesn’t change the fundamental context of the problem, which is that the problem has no objective answer. More about values and purpose of criminal system
What is a massive issue with AI?
Transparency (it’s basically a black box)
Ex. a husband and wife who both applied for an Apple Card — and got widely different credit limits
(even tho shared finances)
How can we regulate AI?
need new laws to regulate artificial intelligence, and some lawmakers are catching up on the issue. There’s a bill that would force companies to check their AI systems for bias
Do our old laws cover AI?
Sometimes!
whether the government is prepared to study and govern this technology, and figure out how existing laws apply
How is there a legal disconnect with developers and law makers?
“You have a group of people that really understand it very well, and that would be technologists,” Turner-Lee cautions, “and a group of people who don’t really understand it at all, or have minimal understanding, and that would be policymakers.”
