Wednesday , June 3 2020
Home / Epsilon Theory / Hateful Memes and Election Season

Hateful Memes and Election Season

Summary:
To receive a free full-text email of The Zeitgeist whenever we publish to the website, please sign up here. You’ll get two or three of these emails every week, and your email will not be shared with anyone. Ever. As the core COVID-19 narrative shifts from the mixture of human and economic tolls that dominated news over the past few weeks to the coverage of lockdown relaxation across the United States, we are starting to observe other topics and language creeping back into the Zeitgeist for the first time. Unsurprisingly, most of them have still been tangentially related to the pandemic. For example, we are observing a lot of shared language in the ether about direct sales and e-commerce (a term I thought had gone out of vogue years ago, to be honest). The framing of these

Topics:
Rusty Guinn considers the following as important: ,

This could be interesting, too:

Ben Hunt writes The Hertz Story Isn’t What You Think

Ben Hunt writes American Idol

Rusty Guinn writes That Old Canard

Ben Hunt writes Through No Fault of Their Own


To receive a free full-text email of The Zeitgeist whenever we publish to the website, please sign up here. You’ll get two or three of these emails every week, and your email will not be shared with anyone. Ever.


Hateful Memes and Election Season

As the core COVID-19 narrative shifts from the mixture of human and economic tolls that dominated news over the past few weeks to the coverage of lockdown relaxation across the United States, we are starting to observe other topics and language creeping back into the Zeitgeist for the first time.

Unsurprisingly, most of them have still been tangentially related to the pandemic.

For example, we are observing a lot of shared language in the ether about direct sales and e-commerce (a term I thought had gone out of vogue years ago, to be honest). The framing of these pieces positions them squarely – and consistently – in context of a response to pandemic-driven buying behaviors.

Likewise, there are fascinating micro-clusters of only-sort-of-coronavirusy language popping out in ways we have not observed before, too. For example, this standard wire service blurb is representative of a wide swath of health care companies reporting, affirming guidance and trading up a bit on the news. It is an odd bit of linguistic uniformity.

And if you can manage to link a holiday like……Mother’s Day with COVID-19, well, then you’ll probably find your way to a brief top billing on our check for language with high interconnectedness across clusters and topics. Nice one, Mike.

We have also observed that story stocks – the ones which have some kind of non-fundamental appeal that crosses typical demographic and investor style boundaries – frequently make their way to the top of the Zeitgeist. It is an unavoidable result of their connection to all sorts of other pop culture, technology, social and cultural trends. Which, in a nutshell, is what we mean when we call something a story stock.


Yet we are observing some emerging consensus topics and language that have little-to-no pandemic relationship, and which we think are likely to be our companions for the next few months. One of the biggest appears to be the role of AI and machine learning in the identification of patterns of online behavior. Or, in the words of researchers at Facebook, ‘hateful memes’.

The article itself is largely a summary of a paper published by the Facebook AI team in connection with their prize-sponsored effort to crowdsource solutions to the problem of systematically analyzing combined text and image data (or other multi-modal forms) that typifies most internet memes. In short, it’s pretty hard to spot whether the most common type of internet meme is abusive, mean or hateful, and we’re still not very good at it. It is technologically fascinating exercise. It is also terrifyingly Orwellian.

To their credit, ZDNet asks some good questions.

A second question was what the scale of the problem is of hate speech at Facebook. Given that the premise of the work is to enlist AI to clean up hateful utterances on social media, it’s important to know things such as how much hate speech is removed on a regular basis, or perhaps a tally to date. Facebook declined to comment.

Facebook says AI has a ways to go to detect nasty memes (ZDNet, May 12, 2020)

M’kay.

There ought to be an impulse to ask “why am I reading this now?” in response to this piece, and perhaps even more so to the underlying Facebook research. It’s reasonable to wonder why its language is so deeply connected to other news being published right now. At risk of answering what is usually intended to be a rhetorical question, I’ll offer what I think based on the language I’m observing in other articles it was connected to: election season.

It’s here, and this is ripe territory for the top-down political narratives we will be discussing at much greater length in the coming weeks. If you have not prepared yourself for a news and social media cycle framed in highly polarized terms of misinformation / censorship / manipulation, now is probably your last chance.

Rusty Guinn
Executive Vice President of Asset Management, Salient. Rusty Guinn is the executive vice president of asset management at Salient. He oversees Salient’s retail and institutional asset management business, including investment teams, products, and strategy. Rusty shares his perspective and experience as an investor on the Epsilon Theory website.

Leave a Reply

Your email address will not be published. Required fields are marked *