Tech Platform Policy Highlights - Digest 6
The Final Scoreboard for Feb: Twitter 5, Meta 4, Microsoft 3, TikTok 3, Apple 2, Google 2, Animoca Brands 1, Bank of Israel 1, China Central Bank 1, GoodRx 1, OpenAI 1, Tinder 1, Twitch 1, Wikipedia 1
This post is part three of a series documenting policy changes and feature improvements introduced by platforms in February 2023. You can find Part 1 and Part 2 here.
TL;DR → In an effort to curb “undesirable” behaviors on the platform:
Instagram India Announces New Policies to Fight Misinformation
Microsoft caps Bing AI chat interactions after “derailments”
A couple of (interesting) tech policies could not be categorized:
Policy changes to curb socially or politically undesirable behaviors on the platforms (such as service denials, discrimination)
Instagram India Announces New Policies to Fight Misinformation
As a part of “Digital Suraksha” Campaign in partnership with India’s Ministry of Electronics and Information Technology, Instagram intends to implement policies such as (a) labeling posts that contain false or misleading information (third-party fact-checked), (b) reducing the distribution of (i.e., removal from Explore and Hashtag pages) posts that are rated as false, partly false, or misleading; (c) blocking or disabling accounts that repeatedly share misinformation or violate Instagram’s community guidelines. These efforts highlight the complexities faced by Instagram in addressing misinformation. Clearly Instagram has figured that the solution requires local collaboration and education (link).
TikTok Launches Account Strikes for Policy Violations
Taking a page out of YouTube’s book, TikTok will implement a strike policy for its bans. Accounts that post content breaking TikTok’s rules will receive strikes and have their offending content removed. The strike system is for minor/inadvertent violations; severe violations such as posting child sexual abuse material, threats of real-world violence, or other extreme content would attract immediate bans. While this simplifies the moderation process for creators, increases transparency and provides opportunity for creators to course correct, it is unclear how TikTok will handle “acceptable” content that did not age well (link).
Microsoft caps Bing AI chat interactions after “Derailments”
GPT-powered Bing chatbot can provide engaging and informative answers, but it can also be manipulated by persistent users who exploit its guardrails. With this realization, Microsoft has limited the count of questions per conversation/session. Initially it was five question-and-answers per individual session but now it has been relaxed up to 20 per session. Microsoft’s action demonstrates the trade-off between its ambitions to overtake Google in search (with the GPT-powered Bing) and the untested, chaotic nature of natural language models (link 1, link 2, link 3).
Twitter tweaks its violent speech policy
Per Verge, Twitter’s new Violent Speech Policy seems to have both broadened and narrowed its scope. Wishing someone harm, for example, was not explicitly banned by the old policy but is by the new one. Policies also seem to have become more vague with regards to protecting marginalized groups: the old policy prohibited threatening violence against “an individual or a group of people”, while the new policy prohibits threatening “others”. With a smaller moderation team, it is unclear how Twitter will enforce these rules with a more vague language (link).
Meta funds tool to help minors erase explicit images online
Meta supports a new platform called “Take It Down”, which allows minors to report and remove their explicit images from the internet. The tool is aimed to help minors regain control over their images and protect themselves from exploitation. This effort reflects a step forward for Meta, which has come under fire in the past for failing to protect minors from abuse. Interestingly, TikTok also participates in “Take It Down”. (link).
Miscellaneous Policy Changes that we could not categorize
At least one of them is geniuinely funny!
Twitter changes algorithm to boost Musk’s tweets
Twitter engineers were apparently asked to alter the algorithm and prioritize Elon Musk’s tweets after his Super Bowl tweet got less engagement than Biden’s. The algorithm now automatically greenlights all of Musk’s tweets and boosts them by a factor of 1,000, resulting in a flood of Musk’s tweets in users’ feeds. The altered algorithm affects all Twitter users and their experience on the platform, demonstrating Musk’s interest in boosting his engagement on Twitter. Of course, Musk is entitled to make any changes on Twitter 😊 (link).
China’s Central Bank revitalizes e-CNY during Lunar New Year
By giving away millions of dollars worth of e-CNY, the digital version of Chinese Yuan, the bank intends to revitalize the use and adoption. The distribution of these tokens happened through various channels such as lotteries, subsidies and vouchers. This also coincides with the bank’s effort to bolster the user experience with the currency by improving the wallet app as well as introducing multiple digital yuan ATMs in Shenzhen. The effort further highlights China’s ambitions to transform into a cashless society. (link)
Research help from Anantesh Mohapatra, Melvin Jabamani, Yiran Liu (Thanks a ton, folks!)