Steal This Idea: Social Media Analyzer

This post is part of my “Steal This Idea” series: free ideas for anyone to take and run with. Learn more about #StealThisIdea.

Content Warning: The inspiration behind this idea was politically-driven. If you’re just here for the ideas and not progressive politics, you may choose to skip the backstory.

After the 2018 mid-term elections proved to be rather disappointing (read: conservative) for the state of Ohio, I toyed with the idea of running for public office. It’s still not something I’ve ruled out, but the emotions it stirred up got me thinking about campaign-oriented technology.

I started thinking about what tools might benefit a younger, progressive candidate and I realized that social media can be a tremendous asset (see: AOC) or a tremendous liability (for example, pretty much every Republican politician).

People my age (early 30s) are the first ones to really enter adulthood in the age of social media; I joined Facebook when it was first rolling out to select Universities, and had been blogging and on MySpace throughout high school.

Unfortunately, when you put yourself into the public eye, you must expect some level of scrutiny. Old Tweets posts will be dug up, tagged photos will be scrutinized, and Facebook posts from your 21st birthday will be surfaced. If you’re not doing this research ahead of time, you can be damn sure your opposition will.

The Idea

The concept is a social media analyzer tool, which ingests posts from a single person across social media channels (Facebook, Twitter, Instagram, et al), including older networks that the target may have used in the past.

These posts would then be analyzed for keywords or topics that might be controversial. Did someone spend a whole lot of time praising someone who they’d rather not be associated with today? Are there long-forgotten blog posts about reckless days of youth, or cringe-worthy, angst-fueled poetry you’d rather not be dug up? Was your early online persona best-described “Edgelord”?

The tool should also look at content being shared or interacted with. Remember the time that someone with access to Ted Cruz’s Twitter account “liked” a pornographic tweet? Is someone regularly sharing content coming from untrustworthy sources like InfoWars or Brietbart?

Beyond finding the content that might be worth cleaning up, the application would also perform sentiment analysis on content. Are you generally a negative person online? Is that a representation of who you are today, and is that something that might turn other people off of you?

Once the tool knows where to look, it could also monitor things moving forward; are you having a bad week and airing dirty laundry in public? While that’s often cathartic and can be seen as “real”, hurling insults on Twitter isn’t exactly great leadership.

Competition

In some initial research, I found that there are some tools similar to this:

• I Flunked My Social Media Background Check. Will You? via Gizmodo
• Social Intelligence operates in this space for employers.
• G2’s list of alternatives to Social Intelligence

Social Intelligence also has a blog post that covers good ways to fail their checks; this could be good training data for any sort of machine learning model.

Monetization Strategies

There are a number of ways this idea could be monetized:

• Charging candidates (and/or people of interest) to see what the system can find on them.
• Charging parties (or potential employers) interested in vetting a candidate.*
• Charging outside parties doing opposition research.*
• In-kind donations to political campaigns.
• Charging candidates searching for themselves, then flagging anything really bad and selling it to their opposition (if you’re even considering this, go read the next section over and over again).

*Ethical Concerns

Given the political inspiration for this service, it’s worth discussing the ethical implications of such a service.

One one hand, the data ultimately comes from the person being researched; if I have a public blog post archived somewhere from high school where I’m making off-color jokes, there’s always the chance that could come up. This is a big part of the “remember: the internet is forever” rule.

On the other hand, a tool that can easily ingest and analyze this data could be abused pretty easily; bad people who have the resources to build/run this service could use it to more effectively discredit more digitally-oriented challengers.

Imagine a well-financed group of evil-doers, holding onto an elected office through undemocratic means like gerrymandering, voter suppression, and the dissemination of outright lies. A tool like this could help find every off-color comment their opposition has ever made online, which could then be exploited — in the absence of standards and ethical electoral practices — turned into a smear campaign.

There’s also something to be said for letting social media be reflections of both who we are and who we were. I’m a very different person at 32 than I was at 22, and I know I’ve [manually] done some clean-up of social media accounts in the past, but these accounts are records of who we’ve been.

Teenage angst, first loves, and early fandoms are reflections of where we’ve come from. Do we want elected officials who have never had a bad day, or angrily penned a late-night post about how a show they loved wrote out their favorite character?

The people running for office are human beings, with human feelings. A tool designed to seek out these “character flaws” has a high potential to be used, if not for nefarious purposes, then at least to further the divide between ideologies.

Could this service make money? Most likely, but I’d warn anyone thinking about building it to consider how it could be used for harm and to build safeguards in from the beginning, including (but not limited to):

• Requiring proof of identity and consent before building a profile for someone.
• Vetting of customers using the tool (e.g. don’t provide service to the bad guys).
• Not licensing the code to parties who don’t explicitly agree to the previous points.

Ultimately, maybe this is a tool that shouldn’t exist. It could be helpful to the next wave of younger candidates, but the potential for abuse is high. If you’re thinking of running with it, I’d strongly recommend thinking through the ethical implications before getting started, and please don’t use it for evil.