Science of Data Ethics (CIS 399) Lectures, Spring 2021 - Shared screen with speaker view
will start in a couple
It’s not clear to me that what is providing any of the moral force to people’s concerns is the thing being “automated” . Rather what I suspect is pushing peoples moral intuitions about these sets of questions is implicit commitments people have to certain normative/moral frameworks (e.g. consequentialist, deontology, virtue ethics, Rawlsian etc). Those are what should be more carefully probed. Does any one these framework make sense any given locale etc?
I’m skeptical of consequentialism....It could be used to justify the slave trade in possible worlds.
I agree that guilt shouldn’t be removed, but in either case, it is not up to the armed forces to decide whether or not to kill - it is an order given to them
So in a way, the person making the decision is already removed from the situation
Very true. Good point.
I respect your opinion, Zach, but I fundamentally disagree with the idea that society would be better if it was fully utilitarian. This mode would not value the elderly, disabled, or sick at all, which is frightening since I believe that ALL humans have inherent dignity.
@Zach I think you probably mean something like the Rawls’ “difference principle “ do what benefits the worst-off members of society of the most and not utilitarianism. Utilitarianism actually can be used to violate all sorts of basic rights.
Zach, you had me convinced with utilitarianism up until you said we should choose 5 over 1 every time. Here the greedy approach doesn't work as intended (i.e. when the 1 person happens to be on the verge of a cute to cancer.)
there's also inherent implementation issues with utilitarian strategies - someone needs to decide how we're assessing value of a particular action, and it's very difficult/impossible to get perfect information about the outcomes of an action
@Gabrielle I definitely understand your point, but think that the conversation of dignity v. fairness gets complicated extremely quickly. In situations with limited resources, all humans do have inherent dignity but you have to take an approach that values more lives.
I know few serious academic philosophers who actually defend utilitarianism in our discipline. Aside from Peter Singer and he admits it justifies killing babies with learning disabilities.
@Alexander I haven’t heard of this principle before, but I think that your definition at face value is roughly what I said. I don’t think I am necessarily focused on just the worst-off members of society, but focusing on a maximization of social welfare. Agreed that utilitarianism for sure has its flaws as we saw in the 1 v 5 where the 1 is about to solve cancer. Obviously it is very tough to calculate the MSB of each person so agreed that this simple framework itself breaks down but has merit in a lot of situations.
@Catherine I think applying the same basic framework to every question doesn’t work but does in the large majority of scenarios. However, we can’t let the perfect be the enemy of the good. It is definitely possible to construct certain hypotheticals where one works more than the other but think as a general idea for making decisions, it can hold up.
@Zach, google John Rawls. He is one of the most famous analytic moral and political philosophers of 20th century. In “Theory of Justice” he characterizes the difference principle and frame’s his framework in opposition to utilitarianism/consequentialism.
Historically in philosophy the first few chapters of this book is known to be one of the most devastating criticism of utilitarianism/consequentialism.
defaulting to utilitarianism unless there are clear reasons to use a different moral framework is also making a moral decision. Stating that constructed hypotheticals are the main enemy of utilitarianism is an oversimplification of the issues it brings in practical scenarios
I would say human nature is the main barrier to utilitarianism
@Adrian good point. In fact, I would argue most ppl do not default to utilitarianism/consequentialism. They default to some version of liberalism. Where they look at how basic rights and they relate to eachother and support eachother. In fact all SCOTUS cases argue on the basis of basic liberal rights not utility.
@Alexander Will do, haven’t read him before. @Adrian I think it’s fine to make a moral decision, our entire society is created on electing people who then create moral decisions for our society. I agree with your point that this is a great oversimplification. @Julia Human nature definitely the main barrier as is bias for the individual (think how many lives would you save in exchange for your own? way higher if thinking about yourself than versus other people)
For sake of clarity by basic liberties I mean rights that are fundamental in that they can only be restricted for the sake of protecting others’ basic liberties. They cannot be restricted for the sake of the general welfare or “utility” or because of the will of a democratic majority, or to increase economic benefits to the less advantaged under the difference principle. They are also inalienable thus cannot be transferred to someone else, unlike property rights. The basic liberties are: Liberty of Conscience; Freedom of Thought and Expression; Freedom of Association; Freedom of the Person & Right to hold Personal Property; Equal Political Liberties; and the Rights and Liberties that guarantee the Rule of Law (due process, fair trial, etc.).
@alexander: definitely a lot of legal precedent in the US for not going utilitarian first, probably for good reason imo. ty for the stuff on rawls@zach: not disagreeing with that specifically. I disagree with the assertion that utilitarianism is the best choice due to its supposed impartiality and objectivity. Even if we only consider the unavailability of perfect information, perfect foresight and impartial decision-making become impossible, and pretending otherwise is making a non-objective decision in favor of utilitarianism
This questionnaire appears quite ridiculous for actually helping conviction of a crime, it makes too many assumptions about what causes someone to become a criminal
is there anything stopping people from lying to make the algorithm think better of them
^good point, Adrian
out of curiosity, is there a place where we can take the survey to find our scores?
not every question is used in the model though right?
Also the administrator asking the questions makes a difference, like whether people feel uncomfortable or scrutinized when having to give their answer
No assumption of measurement error b/w the proxy/measurable variable and the latent/unobserved target variable.