Edition 4 - April 2024

Privacy and AI

Korea’s Personal Information Privacy Commission (PIPC) worked with the Korea Internet & Security Agency to issue guidance to AI platforms after investigating major LLM providers’ services (including Open AI / Google / MS / Meta / Naver / Rutton).

Be sure to subscribe to Asia AI Policy Montior on Substack!

IP and AI

The Japan Patent Office updated its examination handbook with new example cases related to AI inventions. A short explainer slide deck can be found here.

Japan’s Agency for Cultural Affairs Copyright Subcommittee published “Thoughts on AI and Copyright” guidance. It notes the legal grey area of copyright infringement when requesting genAI works “in the style of” rights holders, and notes future work about neighboring rights, moral rights, and other rights to protect voice, and image.

The Nippon Anime and Film Culture Association published comments in response to the Japanese government’s “IP in the AI Era” consultation from the IP Promotion Secretariat of the Cabinet. The comment urges the use of watermarking to identify AI generated infringing material, and urges stronger protection for neighboring rights, to protect recognizable voice actors (e.g. right of publicity).

Hong Kong minister says the city will review its copyright rules in light of AI advances.

Korea’s Ministry of Culture Sports and Tourism - in charge of copyright - issued an English language guide for AI and Copyright (Korean version released last year).

China’s Beijing Internet Court rules that genAI infringes right of personality for a voiceover actor based on the similarity in the genAI output and use of the voice in training data.

Multilateral

Japan’s PM Kishida to announce in May at the OECD meeting a multilateral framework to govern genAI - and promote authentication and origin tools.

Korea and the UK will host the next AI Safety Summit on May 21 and 22 in Seoul. The event will be called the “AI Seoul Summit”.

Korea and the EU issued a statement following the second EU-Korea Digital Partnership Council meeting.

Korea was also welcomed into Horizon Europe - the innovation & research project.

Korea and the US also formed an AI Working Group.

On April 2, leaders from the US and China (Biden and Xi) discussed on a phone call numerous items, including AI-related risk prevention, according to a White House read-out of the conversation. Follow-up meeting by US staff indicates future talks.

China and the African Union issued a joint statement on cooperation on AI, which focused on R&D, talent exchange, and communication channels.

Trust, Security and AI

Microsoft reported that threat actors from China and North Korea have utilized extensive generative AI tools to spread disinformation. The report, titled “Same Targets New Playbook: East Asian Threat Actors Employ Unique Methods” shares recent information on disinformation campaigns fueled by AI with memes, audio, etc.

• What we are thinking: We anticipated in our 2024 Asia AI Policy Forecast that cybersecurity and AI would be a focus for regulators, which is hard and beware: so-called “Fake News” laws can become tools for abuse, as well-documented by the International Center for Not-for-Profit Law.

India’s elections are gearing up and AI will play a role. This brief by Tech Policy Press goes into depth on the issue.

The cybersecurity agencies of Australia and New Zealand joined the US, Canada and UK, to issue best practices on deploying AI safely.

Analysis

This article from Lawfare, suggests that governing AI by governing COMPUTE (the computing power - ie chips) could be an effective way to manage AI. Note the geographic bottleneck on the fabrication side - being Taiwan’s TSMC. Given the outsized importance of Taiwan in AI infrastructure, while also being a target for misuse of AI (see above Microsoft Threat Intelligence Report) - inclusion of the country in global AI governance discourse is imperative.

Meanwhile, the Stanford published the AI Index this month. Regarding Asia, the report noted that although China is behind the US on many AI technologies, there is one area they are leading: AI patent filings.

The report also shows that Asian firms are more concerned about security risks and privacy, as compared to other regions.

The World Economic Forum and Microsoft published AI for Impact: The role of AI in Social Innovation. Looking at Asia - we see social innovators using AI in healthcare and less on environmental issues compared to other regions

Policy Advocacy

Korea PIPC opened public comment until April 30 on seeking case studies and best practices on pseudonymization policies.

China National Information Security Standardization Technical Committee opened comments until June 4 on data security specifications and standards for genAI tools.

New Zealand Privacy Commission opened a public consultation on biometric technology and privacy. Key areas of concern are notice of use, detecting emotions, demographics, attention, health, and other behavior. Note also exceptions on consent, and industry sectors (e.g. marketing).

India’s Competition Commission issued a consultation for a study on AI and Competition in India. Open until June 3.

In the News

The Chinese-owned, Singapore-based TikTok will be forced to sell or shut-down in the US in the coming months (or years depending on litigation).

• What we think: Driven by advanced algorithmic recommendations, this case study will draw the lines on data, security and privacy rules in a cross-border AI-powered economy for the future. Good analysis here (how not to do advocacy in the US) and here (a Cold War casualty?).

An inter-ministry committee in India calls for a whole-of-government approach to AI regulation.

TechRepublic article on the impact to APAC businesses on genAI deepfakes.

Korea announced the formation of a High-Level AI Strategy Council to inlcude 32 members from the public and private sector.

Southeast Asians are going to Taiwan to train for semiconductor jobs.

Tech Policy Press explores AI policy innovations by privacy regulators in Korea, Hong Kong, Australia, New Zealand and Singapore. Written by our editor Seth Hays.

OpenAI opens first office in Asia in Japan to serve local users. Initiatives include an optimized Japanese language offering of the popular genAI chat tool.

Channel News Asia reports on AI use by artists in Southeast Asia. Covers some of the legal questions that face artists using AI under Singapore’s Copyright Law.

Indigenous rights in art, training data and genAI contemplated in this article from Australia about aboriginal art in Adobe’s stock image database.

• What we think: We explored the issue of indigenous data sovereignty and AI at our Taiwan Internet Governance Forum panel in 2023.

Thanks for reading Asia AI Policy Monitor?!

Don’t forget to participate in our work by joining the Asia AI Governance and Policy Center of Excellence (AI ACE) group on LinkedIn here or contact our editor Seth Hays at [email protected].

We are powered by the staff at APAC GATES - a leading non-profit and rights advocacy firm based in Taipei. Check us out at apacgates.com