The push to "fix" algorithms that produce biased results (e.x. hiring algorithms) ignores the very real fact that these algorithms are LITERALLY telling organizations about themselves and their data collection practices.
No, seriously: a hiring algorithm trained on the history of an institution is being trained on the history of choices made BY that institution. When the algorithm spits out a "biased" result, the algorithm isn't wrong: it's just telling the institution about its habits.
Now, clearly these habits reflect the institution's bias, but the instantaneous move to throw the algorithm out, to retool the algorithm, to point to the algorithm as biased, ignores that the algorithm just doesn't fucking work without a database, without training data.
And where does that data come from? Well, the choices made about data collection, the institutional habits used to train it, the history of decisions made by institutions about what is in their best interests, none of which are "neutral" or "fair" or "just."
All this is to say that we should be looking at these algorithms as auditing the institutions they're trained on, as iterating on their habits and the decisions made, and not as simply "biased." And yes, there are some algorithmic systems that are truly biased.
However, we do ourselves a disservice when we just ignore what the algorithms are telling us about the histories of institutions as represented through their datasets. Moreover, we miss what algorithms tell us about the very practice of collecting data.
Long story short, someone give me grant money so I can train hiring algorithms on a decade of hiring data from, say, R1 institutions in America. And then let me run the algorithms on the existing full time employees at those institutions.

I bet those results would be fun.

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Dr. Johnathan Flowers wants to see academia fall.

Dr. Johnathan Flowers wants to see academia fall. Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @shengokai

1 May
Happy May Day! If you’re in a union that represents all faculty at your institution, contingent or otherwise, and you aren’t considering the needs contingent faculty, you fail to understand the purpose of a union.
If you’re at an institution with separate unions for contingent and TT faculty, and you fail to act in solidarity with the contingent faculty union, you are complicit in maintaining the exploitation of your colleagues.
If you’re at an institution with a grad union, and you fail to stand in solidarity with the grad students, you are complicit in maintaining the exploitation of your colleagues and your students.

Worse, if you take on their work duties while they’re striking, you’re a scab.
Read 4 tweets
29 Apr
The GOP's rebuttal to Biden's address was predictable, not just for the content which positions anti-racist work as the "true" source of division, but for its use of a minoritized person to do it. We should expect to see more things like this.
This move is not uncommon: power structures the world over have often use minoritized figures to "sell" their ideology to their allies and would be converts. On this view, one person becomes a metonym for an entire people and demonstrates that "they all aren't alike."
In higher-ed, we see this when administrators find one faculty member, one student, to show how things "aren't so bad," as if that one member can stand in for all members. It is an inversion of their using one member as an example of how all members of a group are "bad."
Read 9 tweets
29 Apr
So, I'm glad Biden named white supremacy as "terrorism" and pointed to systemic racism in law enforcement, but I do want to keep in mind the purposes for doing so. In my mind, Biden has no other choice but to use this language to name the problem.
I'm pointing this out because the actions he's taking do not align with the rhetoric he is using. If white supremacy was as much a terrorist threat as Bin Laden, whose specter he invoked earlier, I would think that Biden would seek to mobilize resources sufficient to the task.
Moreover, his observation would recognize the breadth of white supremacist ideology not simply as a terrorist threat, or as a problem in law enforcement, which organizes our perceptions of how white supremacy works, but as an organizing principle that directs policy.
Read 8 tweets
29 Apr
Me listening to this Biden speech.
Positioning of the need to cultivate an educated workforce to combat China as an existential threat to America?

That certainly won't do anything to stem the tide of violence against AAPI folks.
Touting DARPA like it is a hub of scientific ingenuity and not some engine of the military industrial complex?

Acting as if the space race was some massive scientific endeavor and not driven by nationalism and military interests?

That's some revisionist fucking history there.
Read 4 tweets
27 Apr
It is EXTREMELY risky for early-career people to specialize in LCT philosophies, not simply for the reasons that @BryanVanNorden pointed out in his retweet, but because of the way that departments and the field doesn't invest in LCT philosophies.
An example is how departments advertise for specialists in LCT philosophies. "Non-western," for example, is used as a catchall for "not-anglophone," and even when it is disambiguated into something like "Asian Philosophy," they're not specific as to which Asian Philosophy.
Thus, you have specialists in Chinese traditions competing with specialists in Indian traditions and Japanese traditions all for the same job. And this doesn't get into what happens when ads ask for "non-western" as code for "non-anglophone."
Read 12 tweets
27 Apr
Right, so. Let me talk about the disciplinary implications of Singer's statement here:
@Helenreflects has a good meditation on this at the Philosopher's Cocoon, linked below. I largely agree with Helen's observations, but I thing we should take a broader view: rather than treat Singer's position as being "his" position, we should treat it as the field's position.
To be clear, while Singer's individual ignorance is unfortunate (but unsurprising), it stands to reason that his position is reproduced and institutionalized fieldwide. Singer actually gives us good reason to think this is the case.
Read 15 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!

Follow Us on Twitter!