Welcome to the 43rd edition of Hold the Code! This will be our final official edition for the year but be on the look out for special releases during the summer! Our final edition explains a recent misstep Twitter has made and what laws should exist within the metaverse. Our weekly feature discusses one author's relationship with AI and cringe.
Happy reading!
Twitter Troubles
The Federal Trade Commission (FTC) and the Department of Justice are fining Twitter $150 million for illegally selling users’ data for targeting ads. This follows a similar infraction Twitter faced in December of 2020 for breaking Europe’s GDPR data privacy laws.
Twitter’s Misstep
The FTC found that Twitter breached an order that prohibits companies from mischaracterizing their privacy measures. They claim Twitter has been using users’ email and phone numbers to inform targeting advertisements, under the guise of improving security on their platform.
“As the complaint notes, Twitter obtained data from users on the pretext of harnessing it for security purposes, but then ended up also using the data to target users with ads," said Lina Khan of the FTC.
The FTC reports this practice has affected over 140 million Twitter users.
All about the $$$
Twitter's main source of revenue is from selling advertisement spots. Targeted ads aim to show products or services to specific groups of users based on who is predicted to be most likely to click on or engage with an ad. Targeted ads themselves are not illegal, and many social media platforms use them to make a profit.
However, since at least September 2019, Twitter has been also using user emails and phone numbers to boost targeted ads when users were led to believe this information was only for account security purposes.
In addition to paying the $150 million fine, Twitter must also…
Stop using emails and phone numbers that were illegally collected
Notify users of improper use of security info
Tell users about the FTC’s decision and law enforcement
Explain how to turn off targeted ads
Provide further options for account security that don't need a phone number
Improve their account privacy and security program
Lawless World
Currently, the metaverse is like a lawless space, where chaos rules all. VR space is ungoverned where events such as “virtual gang-rape” can occur, exposing users to extremely inappropirate content. How should this virtual space be governed?
Lawless world
“The metaverse – the blurrily defined term for the next generation of immersive virtual reality technologies – is still in its infancy. But even with crude graphics and sometimes glitchy gameplay, an experience like this can trigger a deeply rooted panic response”
Metaverse spaces can be filled with” hate speech, sexual harassment, paedophilia, and avatars simulating sex in spaces accessible to children” and unfortunately, research has shown that these encounters are far from uncommon.
Currently, Facebook has implemented a “personal boundary” that is a setting that allows for virtual characters to distance themselves from others and hopefully prevent many of these toxic encounters.
Laws of the Metaverse
Though there are not many existing examples of a concrete law for a metaverse, it is believed that the laws of the metaverse would be derived from the laws of physical countries. This could be good for preventing the current toxicity but also problematic due to the fact that issues in the laws of the physical countries would also be present within the metaverse.
Would breaking the law within a virtual space also be equivalent to breaking the law in the physical world? Questions such as these are yet to be answered, but as the metaverse continues to develop, we will begin to see more regulation.
Weekly Feature: When AI Becomes Satisfactorily Cringe
Author (and founder of Google’s Artists + Machine Intelligence program) K Allado-McDowell recently came out with their newest book, Amor Cringe, a novel aimed to explore the self-obsession permeating media culture. An intentionally “cringey” book, Allado-McDowell explains that
“We cringe because we empathize with the embarrassment that others feel, or that we think they ought to feel. Even if it feels bad, this is an intersubjective connection that social media allows us to feel”
While the topic of a TikTok creator stumbling across a grander meaning in life is certainly intriguing, this isn’t the only notable aspect of the book. Namely, Allado-McDowell isn’t the only author, they’re just the only human author. The book was written along with the input of the Generative Pre-trained Transformer 3, commonly called GPT-3.
What is GPT-3?
GPT-3 is a deep learning AI model that works to simulate natural language, and does an amazingly good job of it. It utilizes deep learning, a type of machine learning where instead of being told what to do, the computer is simply given vast amounts of data and told to find its own patterns within it (in this case, it is told to look at the probability distributions from word sequences).
This is part of what makes GPT-3 so well recognized, as the sheer amount of computational time and power, alongside the vast amounts of training data needed to generate such an unsupervised model, means that creating something like GPT-3 can be hugely difficult. Furthermore, as the creator OpenAI wrote, GPT-3 is able to not only create human-like sentences after being given a prompt, it can also be applied to a wide variety of applications such as identifying themes and emotions in writing.
How was it used in Amor Cringe?
Allado-McDowell didn’t want their book to feel like it was a duet with two distinct voices, so in writing the book Allado-McDowell mixed their words alongside GPT-3’s. As Allado-McDowell put it:
“I see GPT-3 (and LMMs in general) as tools that extend human cognition . . . you can compose new texts that you wouldn’t have been able to without the tool. But the tool needs you to activate it.”
Love HTC? ❤️
Follow RAISO (our parent org) on social media for more updates, discussions, and events!
Instagram: @Raisogram
Twitter: @raisotweets
RAISO Website: https://www.raiso.org
Written by Molly Pribble, Dwayne Morgan, and Arielle Michelman.
Edited by Dwayne Morgan