Cyber Week in Review: May 31, 2024
from Net Politics and Digital and Cyberspace Policy Program

Cyber Week in Review: May 31, 2024

"Eyes on Rafah" circulates on social media; China announces $47.5 billion semiconductor fund; Leaked documents detail Google Search algorithms; France lifts TikTok ban in New Caledonia; TikTok releases influence operation report.
Pro-independance demonstrators hold a Kanak flags as they face French riot police during a protest outside the New Caledonian northern province assembly while French President Jacques Chirac was paying a visit to the New Caledonian northern city of Kone o
Pro-independance demonstrators hold a Kanak flags as they face French riot police during a protest outside the New Caledonian northern province assembly while French President Jacques Chirac was paying a visit to the New Caledonian northern city of Kone o Reuters

AI-generated “Eyes on Rafah” image circulates on social media

An AI-generated image has gone viral on social media over the past week. The image, an overview of columns of refugee tents which spell out the words “Eyes on Rafah,” has been reposted nearly forty-five million times on Instagram as of Wednesday morning. Rafah is a city in southern Gaza that has been the focus of an Israeli military campaign since May 6, which has drawn a major outcry from human rights and pro-Palestinian activists. The Eyes on Rafah image is a kind of “softfake,” an AI-generated image used by a political actor to spread their particular political viewpoints, but which is not designed to deceive or mislead in the same way as so-called “deepfakes.” Renee DiResta, technical research manager at the Stanford Internet Observatory, said that the spread of this specific image has advantages in bringing awareness to the situation in Rafah, because it provides a way to call attention to the conflict while not spreading deeply traumatic images of the violence in the city. The relatively tame nature of the image also means that it may be less likely to be flagged by Meta’s algorithms, which have been tuned to downplay political and graphic imagery. The image, like many other forms of softfakes, appears to fall into a gap in Meta’s AI policies, which require “photorealistic video or realistic-sounding audio” to be tagged as AI-generated, but do not necessarily apply to AI-generated art or propaganda pieces. As of Thursday, the image was still appearing on Instagram stories without a label that it is AI generated.

China announces new $47.5 billion dollar semiconductor fund

The Chinese government announced a third round of funding for its largest semiconductor investment fund, the China Integrated Circuit Industry Investment Fund, colloquially known as the “Big Fund.” The new fund, which counts China’s Ministry of Finance as its largest investor, will have nearly $47.5 billion in reserves to draw on to make investments across China’s semiconductor sector; the third round will look to build on the previous two phases of the fund, which were established in 2014 and 2019, respectively. According to analysts, this round will likely focus on areas of chip manufacturing where China has little to no capacity, including the manufacture of high-bandwidth memory chips. The Chinese government has been pushing for years to expand domestic microchip production and has accelerated that effort as the United States has begun to clamp down on Chinese businesses’ ability to buy advanced microchips and microchip manufacturing equipment from firms in Taiwan, the Netherlands, and Japan.

Documents leaked from Google outlining its Google Search algorithm

More on:

Artificial Intelligence (AI)

Cybersecurity

Digital Policy

An individual named Erfan Azimi leaked more than 2,500 pages of technical documents that detail the application programming interface (API) documentation from Google’s Search division earlier this week. Google’s search algorithms have been among the most protected aspects of the company’s internal systems for years; an entire cottage industry has arisen around so-called search engine optimization (SEO), which aims to maximize the number of visits to a particular website by ensuring a site appears high on a search engine’s results. The leak outlines a number of different aspects of how Google set up its search function, including what data Google collects from webpages, sites, and searches; although it doesn’t reveal the most sensitive parts of Google’s search algorithm, such as the weight of particular parts of its search algorithm. The leak may have originated from GitHub, and may have been exposed when the documents and internal corporate pages were briefly made public sometime between March and May 2024.

French government lifts TikTok ban in New Caledonia

The French government has lifted a ban on TikTok in the territory of New Caledonia, which was put in place after a state of emergency was declared in the territory on May 16 in the wake of social unrest. The French government has deployed more than 3500 security personnel to the island in the wake of the unrest, as indigenous Kanak people have been protesting against a series of electoral reforms; the reforms would grant voting rights to French residents who have lived in the territory for more than ten years on the island, which some Kanaks feel would dilute their electoral power on the island. The French government had justified the ban on TikTok by alleging that it was being used to spread disinformation, incitement, hatred, and plans to incite violent protests on the island. The French government also accused the Chinese and Azerbaijani governments of interfering in the crisis, which both governments denied. Despite the end of the state of emergency, the French government is still imposing a curfew in the territory, and has banned public gatherings, the public carrying of weapons, and the sale of alcohol in the territory.

TikTok releases a report covering influence operations on the platform

TikTok released a report last week that describes the influence campaigns that were shut down on the platform between January and April of this year. In all, the company stopped 15 campaigns which were utilizing more than 3,000 accounts worldwide, with the largest campaign being a pro-Ukrainian campaign that reached nearly 2.6 million people. The report primarily describes campaigns run by individual governments focused on influencing their respective citizens, including in Equator, Guatemala, Serbia, and Venezuela; however, it also describes some directed at external groups, including an Iranian government-backed campaign aimed at spreading pro-Iranian views in the United States. TikTok also described the process by which it evaluates and counters covert influence operations, which are defined as “networks of accounts strategically [working] together to mislead people or our systems and influence public discussion.” TikTok says that it primarily looks for evidence that accounts are coordinating with each other, that they are aiming to mislead systems or users, and that they are attempting to mislead or corrupt public debate, opinions, or beliefs.

More on:

Artificial Intelligence (AI)

Cybersecurity

Digital Policy

Creative Commons
Creative Commons: Some rights reserved.
Close
This work is licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) License.
View License Detail