Key AI developments in March 2024

Keeping up with developments in AI legislation, regulation, investigations, penalties, legal action, and incidents is no easy task, which is why we have created Tracked, a bite size version of our global AI Tracker. Each month, Holistic AI’s policy team provides you with a roundup of the key responsible AI developments from the past month around the world to keep you up to date with the ever-evolving landscape.?

Europe

1. EU AI Act Passed by the European Parliament

• The European Parliament approved the landmark EU AI Act on 13 March. The progress marks the first of two stamps needed before the Act is officially adopted.
• The Council of Europe will then have to approve the Act, which is set for May this year.
• The legislation takes a risk-based approach to regulating AI, with systems subject to varying levels of obligations according to their risk classifications. High-risk applications have the most stringent obligations, while systems deemed to have an unacceptable level of risk are banned altogether.
• The application of the Act's provisions will be phased, affecting various sectors at different stages.

2. Italian Data Protection Authority opens an investigation into OpenAI’s Sora

• Italy’s data protection authority Garante opened an investigation into the AI developer’s text-to-video service in early March.
• While the investigation did not detail any specific allegation, Garante stated that it was investigating the implications of Sora on the use of personal data.
• The investigation comes in the wake of Garante’s ban of ChatGPT last year over alleged breaches of the Union’s privacy rules.

3. European Commission publishes guidelines for mitigating risks ahead of elections under the DSA

• On 26 March, the European Commission published guidelines for Very Large Online Platforms (VLOPs) and Very Large Online Search Engines (VLOSEs) under the Digital Services Act (DSA) ahead of the June European Parliament elections.
• The guidelines outline best practices for the mitigation of risks before, during, and after electoral events with a view to safeguarding fundamental rights, such as the right to freedom of expression.
• Recommendations include reinforcing internal processes such as establishing dedicated internal teams, implementing tailored election-specific risk mitigation measures, adopting specific measures for generative AI, establishing an incident response mechanism, and monitoring the effectiveness of measures.

4. EU’s Digital Markets Act Comes into Effect

• The DMA, the European Union’s legislation that addresses fair competition in the digital economy, came into effect on 7 March 2024.
• Its entry into force is set to see major tech companies face intense scrutiny and potential fines for anticompetitive behavior.
• The Act allows the EU to issue heavy penalties of as much as 10% of a company’s total annual worldwide revenue, and up to 20% for firms who repeatedly flout the rules.
• Alphabet, Amazon, Apple, ByteDance, Meta, and Microsoft have all received the “gatekeeper” designation and are thus required to make changes to reduce anticompetitive behavior. ?
• Apple will now have to allow other app stores onto iOS, Meta’s WhatsApp will have to be interoperable with other messaging services, and Google must let EU users choose which services share their data.

5. Italian Competition Authority imposes 10 million Euro fine on TikTok

• Italy’s Competition Authority the AGCM has fined ByteDance’s most popular app for failing to sufficiently protect minors.
• AGCM accused TikTok of failing to adhere to its own guidelines that claim it ensures a “safe space.”
• The company issued a statement disagreeing with the decision and defending its track record.

6. European Commission opens formal proceedings against AliExpress

• The EU Commission has opened proceedings against the online retail giant to assess whether it has failed to comply with the Digital Services Act (DSA).
• The proceedings will verify whether the company has enforced policies regarding consumer health, illegal content, intentional user manipulation, and the promotion of illegal or harmful products.
• AliExpress, which is owned by Chinese company Alibaba, is considered a Very Large Online Platform (VLOP) under the DSA.

7. EU AI Office hiring

• As the EU AI Act moves to the next stage of approval, the EU Commission has released two new roles in its AI Office.
• The Office, launched at the end of February, announced that it was specifically looking for technology specialists and administrative assistants.
• Technology specialists will help enforce and supervise new rules for generative AI models.
• Administrative assistants will be responsible for supporting the operations and enforcement tasks of the AI Office, among other administrative duties.

US

8. Federal AI CONSENT Act introduced

• On 19 March 2024, Senators Ben Ray Luján and Peter Welch announced the introduction of the Artificial Intelligence Consumer Opt-In, Notification Standards, and Ethical Norms for Training (AI CONSENT) Act to require companies to obtain consent from individuals if their data is used to train AI models.
• If passed, the FTC will develop regulations for the provision of clear and conspicuous disclosures of data use as well as what constitutes express informed consent and mechanisms for revoking consent.
• The FTC would also be directed to conduct a study on data de-identification methods for the training of AI models.

9. New York City’s Technology Office announces progress on AI Action Plan

• New York City’s Office of Technological Innovation (OTI) announced progress on implementing its AI Action Plan, which aims to guide the responsible effective deployment of AI technologies that impact New Yorkers.
• Two key initiatives have been established: the AI Steering Committee and the AI Advisory Network.
• The AI Steering Committee is chaired by the city’s Chief Technology Officer (CTO) Matthew Fraser and includes 23 representatives from 16 city agencies.
• The AI Advisory Network includes subject matter experts from industry, academia, labor, and civic organizations. The group will focus on advancing core initiatives around designing and implementing a robust governance framework and building external partnerships.
• The OTI also published guiding principles for responsible AI use by city agencies, covering areas like appropriateness, reliability, fairness, privacy, cybersecurity, and transparency.
• City leaders from various municipal agencies expressed support for the initiatives.

10. Rhode Island introduces law to prohibit AI in determining rent

• A trio of Rhode Island state representatives introduced House Bill 8058 on 13 March 2024 to prohibit the use of algorithms or AI to set residential rental amounts and landlord price fixing of residential rents.
• The Bill states that violations would constitute a deceptive trade practice and violators are subject to civil suit.
• The House Judiciary Committee is now in the process of reviewing the Bill.

11. Bill establishing New York AI Ethics Commission introduced

• New York State Senate Bill S88755 to establish the New York Artificial Intelligence Ethics Commission was introduced and referred to the Committee on Internet and Technology on 7 March 2024
• The Commission would act as a regulatory body to ensure the ethical use of AI within New York through curated ethical guidelines, project reviews, and the provision of education resources
• The Commission would also be tasked with advising the legislature and investigating complaints about unethical AI practices, and would have authority over both AI systems used by state agencies and private entities

12. OpenAI copyright infringement cases consolidated

• As of 13 March 2024, the slew of separate copyright cases against OpenAI in New York and California could potentially become consolidated due to their similarities.
• Some high-profile separate cases involving Paul Tremblay, Sarah Silverman, Richard Kadrey and Chabon have already been consolidated.
• These authors have now requested that the New York Times case against OpenAI be dismissed and brought into their own case.

13. Federal AI Governance and Transparency Act introduced in House Oversight and Accountability Committee

• H.R. 7532 the Federal AI Governance and Transparency Act was introduced to the House of Representatives by Committee Chair James Comer (R-KY) and Ranking Member Jamie Raskin (D-MD) on 5 March 2024.
• The Act states that the use of AI in the federal government is permissible if it does not disproportionately impact certain groups negatively or disproportionately favor certain groups.
• The Act focuses on public procurement of AI systems and specifically covers the design, development, acquisition, use, management, and oversight of AI within the federal government.
• H.R. 7532 calls on the Office of Management and Budget to issue guidance on the measures that providers must meet to be contracted or collaborate with government agencies.

14. Florida S1680 on Advanced Technology passed

• On 6 March 2024, The Florida legislature passed Senate Bill 1680, which created the Government Technology Modernization Council.
• The Council would study and monitor the development and deployment of new technologies.
• The Bill also requires the Council to provide reports on recommendations for procurement and regulation of such systems to the Florida Governor, the President of the Florida Senate, and the Speaker of the Florida House.

15. Senate Bill 150 for Artificial Intelligence and Cybersecurity signed into Law in Indiana

• On 13 March? 2024, Indiana Governor Eric Holcomb (R) signed into law Senate Bill 150 for Artificial Intelligence and Cybersecurity after its adoption on 7 March.
• The law creates a temporary AI task force to serve the General Assembly by conducting a study on the potential use and risk of AI in state agencies.
• SB requires public entities to regularly check and prove their cybersecurity measures when connecting to the state's tech infrastructure. The law also stipulates specific rules for compliance, domain names, and user authentication.
• It also creates new requirements for state agencies, such as listing their AI systems and the details regarding vendors, functions, security, costs, and potential biases.?

16. NVIDIA sued for copyright infringement

• AI and copyright clashes continue as three authors sued Nvidia for allegedly using their books without permission to train its NeMo AI platform.
• The authors Brian Keene, Abdi Nazemian and Stewart O'Nan argue that the dominant chips manufacturer should pay damages and destroy all copies of the Books3 dataset used to power NeMo LLMs.
• The Book3 dataset was reportedly copied from the BiblioTek, a library of pirated books that was first shared by Hugging Face and then later taken down. ?

Global

17. UN adopts first global resolution on AI

• The United Nations General Assembly unanimously passed the first global resolution on AI on 21 March 2024.
• The non-binding resolution recommends countries safeguard human rights, protect personal data, and monitor AI for risks.
• It marks a striking moment of unity among major countries, with both the United States and China co-sponsoring the resolution.
• The resolution is one of many voluntary initiatives passed by multilateral bodies and governments around the world concerning AI.

18. Indian government issues then backtracks requirements for unreliable AI models

• On 1 March 2024, India’s Ministry of Electronics and Information Technology (MeitY) announced that all “under-testing” or “unreliable” AI models be given permission from the Indian government before they can be made available in India
• On 15 March, however, India revised its earlier statement. Instead, “untested” or “unreliable” models would need to come with a disclaimer that indicates their responses may be unreliable.
• The latest announcement comes as the country is preparing for its national elections this summer.
• It is the latest development in the government’s fluctuating stance on AI; in 2023, the government indicated a laissez-faire approach to the technology.

19. Singapore Data Protection Authority Issues AI Guidance

• On 1 March 2024, Singapore’s Personal Data Protection Commission (PDPC) issued the Proposed Advisory Guidelines on Use of Personal Data in AI Recommendation and Decision Systems (the Proposed AI Advisory Guidelines) for public consultation.
• The Guidance aims to clarify the use of personal data to train and develop AI systems, information to be provided to consumers for obtaining lawful consent, information for third-party developers employing AI models and best practices for complying with the Personal Data Protection Act of 2012.
• It is organized into three broad categories of the implementation of AI systems: development, deployment, and procurement.

20. Japan to finalize guidance on AI and Copyright

• After soliciting public comments on its draft “Approach to AI and Copyright”, Guidance, Japan’s Agency for Cultural Affairs (ACA) has issued the latest draft before finalizing the text later this month.
• The approach appears to endorse a broad allowance for AI developers to use copyrighted work in their training data.
• In general, the country has chosen not to adopt any prescriptive laws on AI and developers must instead rely on relevant existing legislation, such as those on data protection.

Holistic AI Policy Updates

We recently hosted our monthly Policy Hour Webinar with Kai Zenner, Head of Office & Digital Policy Adviser to MEP Axel Voss to discuss how the text will be applied in practice now that the countdown to compliance has begun. Give it a watch here.

We’re also proud to have participated in the finalization of the Framework Convention on AI as part of our Observer Status with the Council of Europe.

Want to dive in?

Check out our blog for deeper insights on key AI developments around the world from our policy team.

Authored by Holistic AI’s Policy Team.