Hub News
On December 7 and 8, the Hub hosted a conference on Data Governance in the Age of Generative AI. Recordings of all the sessions have been posted on our YouTube page.
Today at 11 AM EST, the Hub will host Professor Baobao Zhang of Syracuse University to discuss how she managed to seek, inform and involve the public in the governance of AI. Dr. Zhang will describe the process of convening a public assembly and what the group recommended, and then the audience can ask Dr. Zhang questions about her process and findings.
|
|
|
Susan Aaronson wrote an op ed on the implications of the US retreat on digital trade in Fortune.
On December 20, Aaronson lectured at the UK PTO and Chatham House and on December 21 at C4AI, University of Sao Paolo, Brazil, on her paper Data Disquiet: How Generative AI is Changing Data Governance.
|
|
|
Digital Trade and Data Governance News
|
-
South Africa proposed terminating the World Trade Organization moratorium on e-commerce duties at the upcoming 13th ministerial conference and establishing a fund to help developing and least-developed countries “address the digital divide.”
-
The EU’s Data Act entered into force.
-
The EU’s Digital Markets Act is already forcing major platforms to change how they operate in the EU. Google and Meta have announced changes for EU users including new browser choices and allowing users to de-link services.
|
Artificial Intelligence News
|
-
The European Union has reached a political agreement on the landmark AI Act. Read the final text here; find more information about it here; see the new governance structures that will be set up here; and read about the French government’s continued issues with the Act here.
- The US National Science Foundation opened its 2-year pilot of the National AI Research Resource or NAIRR.
-
The Federal Trade Commission initiated an inquiry into generative AI investments and partnerships, issuing 6(b) orders to Alphabet, Amazon, Anthropic, Microsoft, and OpenAI. The FTC’s investigation will focus on aspects of corporate relationships such as strategic rationale, competitive impact, and information sharing related to AI development, with the intent being to ensure there are no distortions of innovation and competition.
- OpenAI is forming a Collective Alignment team to integrate public input into its AI models and align them with human values.
-
Singapore’s Infocomm Media Development Authority published a Model AI Governance Framework for Generative AI.
-
Meta and IBM launched the AI Alliance to advocate for an open-science approach to AI development.
-
Beijing-based startup 01.AI, led by Kai-Fu Lee, a former executive at Google and Microsoft, released an open-source AI model, Yi-34B, that surpasses Meta's Llama 2 in performance on various benchmarks.The UN Secretary-General’s AI Advisory Body released its Interim Report, “Governing AI for Humanity.”
- Rwanda released its first National Artificial Intelligence Policy.
- A bipartisan group of US Representatives created an AI Working Group.
|
-
The Aapti Institute, Connected by Data, Datasphere Initiative, and MyData Global released a collaborative publication “In this together: combining individual and collective strategies to confront data power.” The paper emphasizes the complementary nature of individual and collective approaches in addressing challenges within the current data governance status quo, aiming to add clarity and complementarity to the advocacy efforts of non-profit organizations.
- Researchers discovered that large language models (LLMs) can exhibit persistent deceptive behavior, even evading standard safety training techniques.
-
Ashique KhudaBukhsh’s research unveils vulnerabilities in Google's PaLM 2 LLM, showcasing instances of disturbing toxic content bypassing guardrails.
-
Dr. Sarah Mercer introduced Willowbrook, a simulated society powered by LLMs like GPT-4, in a demonstration of their advanced non-linguistic cognitive abilities. The simulation showcases LLMs’ capacity for sophisticated content generation, complex interaction simulation, and nuanced conversations, alongside challenges in pseudo-reasoning, attention issues, and linguistic implicature struggles.
-
Researchers from the University of Bologna and European University Viadrina Frankfurt explore the legal and regulatory challenges of generative AI in the EU. Their research scrutinizes aspects of liability, privacy, intellectual property, and cybersecurity, evaluating the adequacy of existing and proposed EU legislation including the AI Act.
- Aurelia Tamò-Larrieux, Clement Guitton, Simon Mayer, and Christoph Lutz explore the role of regulation in establishing trust in AI. The study suggests a nuanced framework for policymakers, distinguishing between regulatory strategies that significantly impact trust in AI and those with more limited influence.
- George Mason University researchers Joy Buchanan and William Hickman find that individuals tend to trust content believed to be human-written more than AI-generated statements.
-
David Evan Harris advocates for stringent regulations on open-source AI systems, emphasizing the responsibility of developers and deployers for potential societal and democratic threats.
-
A recent Consumer Reports study reveals alarming levels of online surveillance, showing that 186,892 companies shared user data with Facebook, averaging 2,230 companies per participant.
-
A recent McKinsey report on the effects of generative AI on Black communities highlights its potential global economic impact but raises concerns that it could widen the racial wealth gap in the United States by $43 billion annually.
|
Opportunities for Public Comment
|
Opportunities for Public Comment
|
|
|
Manage your preferences | Opt Out using TrueRemove™
Got this as a forward? Sign up to receive our future emails.
View this email online.
|
1957 E St NW None | Washington D.C., DC 20052 US
|
|
|
This email was sent to .
To continue receiving our emails, add us to your address book.
|
|
|
|