Cyber Week in Review: May 24, 2024
from Net Politics and Digital and Cyberspace Policy Program

Cyber Week in Review: May 24, 2024

UK and South Korea host AI Safety Summit; China rolls out Xi Jinping Thought chatbot; House committee passes bill restricting export of AI systems; Biden meets with Kenyan President; TikTok to limit state-backed media.
Han Duck-soo, South Korean Prime Minister, gives a speech during the opening ceremony of the AI Global Forum in Seoul, South Korea on May 22, 2024
Han Duck-soo, South Korean Prime Minister, gives a speech during the opening ceremony of the AI Global Forum in Seoul, South Korea on May 22, 2024 Kim Soo-Hyeon/Reuters

UK and South Korea host second AI Safety Summit in Seoul

The South Korean and UK governments cohosted the second convening of the AI Safety Summit in Seoul, South Korea this week. The event took place over two days, with the first day consisting of a virtual meeting between South Korean President Yoon Suk Yeol and UK Prime Minister Rishi Sunak, and the second an in-person meeting of digital ministers and ambassadors from a variety of nations and civil society, industry representatives, and members of civil society. There were several prominent announcements around the summit, with the U.S. Commerce Department launching an international consortium of AI Safety Institutes on day one. The new consortium will build on the cooperative framework between the U.S. and UK AI Safety Institutes, which was unveiled at the first AI Safety Summit in Bletchley Park in November 2023. Secretary of Commerce Gina Raimondo said that the United States, UK, Japan, Canada, Singapore, and the European AI Office would all participate in the new consortium, which will likely be expanded to include other nations. The goal of the group is to promote and advance AI safety, and create a network of institutions, communities, and coordination around the new developments in AI safety. Several companies, including Google, Meta, and OpenAI, pledged to adhere to new voluntary commitments as part of the summit, including agreeing to shut down their most advanced models if the companies can’t rein in certain risks; however, the companies did not specify what exactly these extreme risks would entail. Unlike the first meeting of the AI Safety Summit, the second edition had a much lighter civil society presence, with most of non-governmental presence made up of large foundation model developers.

 

China rolls out chatbot based on Xi Jinping Thought

The Cyberspace Administration of China (CAC) has rolled out a new chatbot that provides answers based on the Chinese President Xi Jinping’s political philosophy, known as Xi Jinping Thought. The model is currently only being used at a research center affiliated with the CAC, although it may be released to the broader public in the future. The model is trained on a variety of different data sources, including existing documents created by the CAC, books written in Xi Jinping’s name, and other data sources. The model is apparently capable of answering questions, creating reports, summarizing information from external sources, and translating between Chinese and English. The CAC has tried to exert more control over the development of generative AI models in China over the past two years, issuing requirements that developers ensure models match “core socialist values,” and  do not output content that “subverts state power.” Developers have faced several restrictions, including requirements that they only train models on a state-approved dataset, as well as significant penalties for failing to ensure their systems meet government standards. Last year, a model built by  Chinese AI developer iFlytek output passages that criticized Mao Zedong, which led to a 10 percent decline in the company’s share price and prompted the founder to issue a public apology.

 

House Foreign Affairs Committee votes to advance bill restricting export of some AI systems

More on:

Cybersecurity

Artificial Intelligence (AI)

China

Digital Policy

The House Foreign Affairs Committee voted to pass a bill that would allow the Biden administration to restrict the export of certain AI systems. The bill is largely aimed at preventing the export of advanced systems to China and passed with support from both Democrats and Republicans. The bill expands the Commerce Department’s Bureau of Industry and Standards (BIS) authority to restrict the export of so-called “dual-use systems” to include AI systems. Those export authorities, which were originally created under the Export Control Reform Act of 2018, have previously been used to restrict the export of certain kinds of semiconductors to China. According to one of the bill’s sponsors, House Foreign Affairs Committee Chairman Michael McCaul (R-TX), the bill will mainly be used to enforce export controls on closed source models, which do not have publicly available code, rather than open-source models.

 

Biden meets with Kenyan President William Ruto and announces technology investments

U.S. President Joe Biden hosted Kenyan President William Ruto for a state visit this week, and announced that the United States will designate Kenya as a major non-NATO ally, giving the country access to additional aid and security guarantees. Biden also outlined a series of investments in technology throughout Kenya, including a $1 billion investment in a cloud computing data center in East Africa by Microsoft and United Arab Emirates-based AI firm G42; a nearly $40 million pledge to build democracy, human rights, and governance initiatives in Kenya that will seek to encourage women’s political participation and leadership and advance digital democracy; and a new EDTECH Africa initiative in partnership with Microsoft, Mastercard’s Center for Inclusive Growth, and several historically black colleges and universities (HBCUs) from the United States. The visit reflects expanding ties between the two countries, with the United States increasingly looking to Kenya as a partner to push back against China’s expanding influence in the region.

 

TikTok announces it will limit Russian and Chinese state media on its platform

TikTok announced that it was limiting the spread of videos from Russian and Chinese state-affiliated media accounts, and preventing these same companies from advertising, ahead of the 2024 U.S. election. Previously, the company had only labelled such posts and ads as state-sponsored content, rather than restricting their reach algorithmically. The change will apply outside of media outlets’ homes countries, meaning that a post from the People’s Daily, the official newspaper of the Chinese Communist Party, would see its reach restricted in any country other than China. TikTok’s peers in the social media space have previously taken action to restrict the spread of state-sponsored content online, with both Facebook and YouTube announcing they were blocking Russian state media from running ads on their platform; suspending the reach of several Russian state-owned YouTube channels; and preventing them from monetizing their platforms, after the 2022 Russian invasion of Ukraine.  TikTok has faced major headwinds in the United States since President Biden signed an omnibus bill last month that will force TikTok’s Chinese parent company, ByteDance, to either divest from TikTok, or see it banned in the United States.

More on:

Cybersecurity

Artificial Intelligence (AI)

China

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