Algorithmic Gatekeeping, What do we think? (15-25 min)
Scary stuff? Overreach for this kind of thing?
On a small-scale level this may not be a big deal or impact you in anyway. An example would be filtering your social media feeds so that you are seeing funny dog clips. I think everyone would be fine with this type of manipulation. I believe when it comes to more serious news and information; algorithmic gatekeeping can be dangerous. For a platform to be able to manipulate these algorithms, they can almost control what you think and believe in. When I am on social media, I usually will not search for information rather just read what is there when I open it. If I just believe and trust the first few things I read as facts, these algorithms are manipulating me into a certain way of thinking.
Useful metaphor?:
Facebook is the lens in which we view our social media lives, we are subject to Facebook bias, and we are blind to Facebook’s control. We may be the head, but Facebook is the neck.
Acceptable in some cases?
Personally I don’t think that is an invasion of privacy I believe its is a strategics and smart business move by that specific company. Sure it can be a little misleading and manipulative and could cause me to go back to the website and buy the clothing that I wasn’t planning on getting before.
Of course there are always loopholes in these rules. There are always ways to make this harmless algorithm into a much more controversial topic like the election. The question is though should the small group that abuses this tool ruin it for everybody who is playing fair? Or should we try to enforce stricter and more specific guidelines for the use of algorithms. While you believe that making these algorithms illegal is the answer I believe stricter regulation could be the middle ground which would leave the harmless groups unscathed.
Don’t we need a filter with so much information? But should we still worry about that?
One line that stuck out to me most in the reading when comparing the use of algorithms for standard data analysis with the use of things like Facebook algorithms was “In subjective decision-making, there is no such “correct” answer with which to anchor and evaluate the algorithm’s operations.” When it comes to analyzing normal data sets for economic or sales purposes it may be necessary to filter the data, but when it has to do with people’s opinions and understanding, filtration must be taken very seriously. The information that is made readily available to people through social media can have a major impact on the opinions they develop on current events, politics, and much more.
How do we respond as writers? Let’s try this simulation (30-45 min)
Go to my GitHub respository.
At the bottom of the page you should see a button that looks like this:
Click that button. Be patient, it takes some time to load the Jupyter Notebook that hosts the newsfeed simulation. (NOTE: When I tested this out, it worked in both Chrome and Firefox, but for some reason did not load in Microsoft Edge).
Now, click the Jupyter Notebook (first option, see below):
Follow the instructions at the top of each box that follow the #. The only time you click is to make sure your cursor is in the relevant box and then click the button at the top that looks like this:
Most boxes will prompt you to type something in. Type it in and then hit enter.
When you get to the very last box, it will print out some discussion questions that we will talk about as a group.
So what do we do when the documents we produce may circulate into these environments? (20-30 min)
What can you do when you compose elements of these documents? How do you think about how the document may appear as a post that depends on engagement from readers in a certain way that is different in a newsfeed compared to, say, a television program or print newspaper?
Some bloggers this week had some ideas on what to do, but I think there is even more to think about then what is below:
1.
At this time, it is hard to determine how many people will see your content. Put a cheery spin on a social justice issue and you might just sneak past the algorithm.
2.
With the use of algorithmic gatekeeping, my campaign of structurally deficient bridges could be shown on news feeds of people who were never looking for it. People who have liked certain articles pertaining somewhat to this topic could have my campaign pop up for that person to read. I believe gatekeeping would restrict my campaign more than help. I don’t think there are many people who are interested in this topic, making it more difficult for my campaign to be filtered into many people’s timelines.
3.
On Facebook, a user can “sponsor” an item to show up on people’s news feeds with similar interest, proximity to the creator, or other factors. In my group’s campaign, creating a sponsored post would be beneficial to promoting our event in which a presentation will be held. However, it would be important to take into account Tufekci’s term of algorithmic gate in order to determine who exactly the post would reach. This would be essential in gaining the right attention and crowd for our event.
Going back to the Target example the author used, it would be necessary to make sure that the event goers were mostly commuters, or simply people that travel across one of Pittsburgh’s three rivers almost every day (as our campaign is concerning the poor structure of Pittsburgh’s bridges). This would mean that the people that are most likely concerned or affected by the goals of our campaign our getting reached. It would not be efficient to be advertising to, for example, people who live on the South Side and also work on the South Side.
4.
What else? What about how you write headlines? Or the images you include for the thumbnail? Teaser text?
This takes me back to our first reading of the semester. Is there a worry about “rhetorical stance” balance in Booth’s terms? Remember his comment about the title of a book being surveyed by 200 people and his distaste for that? Is that a problem here?
Be one with the code, what else is possible for how your writing reaches the world? (15-30 min)
Beside clicks, likes, comments, and shares, what are some other ways Facebook could use what you do to create a model of who you are? To help keep you engaged to feed you ads? (If time), let’s try to write some pseudo-code of another algorithmic gatekeeping method in addition or in place of what I have in my simulation.
Other examples of algorithmic gatekeeping beyond Facebook? What about Twitter (e.g., “In Case You Missed It”, trends)? Other platforms? What about search engines? (e.g., test this by going to a public computer, opening your favorite search engine, and compare the results you get on your personal device and the public computer that has no association with you).
Break (15 min)
Let’s workshop a piece that has communities they want to reach in mind in a digital space (30-45 min)
kemeducation.tumblr.com
I have to be honest, I don’t know a lot about Tumblr, but from what I could gather, I thought there was some cool attention paid to two things:
- Using the affordances of the medium to make available for circulation material that would be interesting and engaging to their target audience.
- There is a tonal consistency across posts that would be appealing for repeat engagements to follow this blog on a regular basis
Some questions to think about as you read:
- What do you think about my first point here? Do you see evidence of that in any one post or two? What is it? Be specific? Or, am I off the mark? Please challenge this point as you see fit.
- What are some examples of tonal consistency across posts? What does the writer do with language and images to do this? What are some specific examples?
- What might be some future posts they do? What are some existing posts that could go in other directions if revised?
Finally, does the first half of class relate to this in any way? How do we think of social media environments and consequences for circulation of this public writing?
Work on campaign plans (15-20 min)
Take some time in class to start preparing for presentations on 11-29. I’ll come around for questions.
Admin (5-15 min)
Things to work on:
-Pedagogical Object. Stay in touch with group, you should know what you are doing and will have already started by the time I see you next. The proposal at end of class on 11-29, I’m not looking to “approve” what you do, as much as provide some feedback on some things to think about. Do what you think teaches, make sure you include your group’s own writing as part of that.
-Start thinking about which campaign piece you want to revise for final. Choose something you are excited to work on rather than something you think is either “really good already” or “really bad and I want to make better.” I want to see work that you take in a new direction or refine in such a way to build on its direction that you are excited about.
-Campaign plan work some more.