With the EU's AI Act, the experience of open-source software can provide guidance for the regulatory approach to open machine learning systems.
The AI Act is seen as a global precedent aimed at addressing risks associated with artificial intelligence while encouraging the development of cutting-edge technologies. One key aspect of the Act is the support it provides for open-source, non-profit, and academic research and development in the AI ecosystem. Such support ensures the development of safe, transparent, and responsible AI systems that benefit all EU citizens.
Drawing from the success of open-source software development, policymakers can formulate regulations that encourage open AI development and protect user interests. By providing exemptions and proportionality requirements for open machine learning systems, the EU can promote innovation and competition in the AI market while maintaining a thriving open-source ecosystem.
Several organizations representing commercial and non-profit stakeholders, including GitHub, Hugging Face, EleutherAI, and Creative Commons, have jointly released a policy document calling on EU policymakers to protect open-source innovation.
These organizations have put forward five recommendations:
1. Clearly define AI components: Clearly defining AI components will help stakeholders understand their roles and responsibilities, promoting collaboration and innovation in open ecosystems.
2. Clarify exemptions for collaborative development of open-source AI components: To encourage open development, the Act should explicitly state that contributors to public repositories are not subject to the same regulatory requirements as commercial entities.
3. Support coordination between the AI Office and the open-source ecosystem: The Act should encourage inclusive governance and collaboration between the AI Office and open-source developers to promote transparency and knowledge exchange.
4. Ensure practical and effective research and development exceptions: By considering the provisions of Article 2(5d) in conjunction with the Council and Parliament, limited practical testing should be allowed to facilitate research and development without compromising safety and responsibility.
5. Set proportionality requirements for "base models": Differentiate requirements for base models based on their usage and development models, including open-source approaches, to ensure fairness and promote competition.
Open-source AI development offers several advantages, including transparency, inclusivity, and modularity. It allows stakeholders to collaborate and build upon each other's work, resulting in more powerful and diverse AI models. For example, the EleutherAI community has become a leading open-source machine learning laboratory, releasing pre-trained models and code repositories that promote foundational research and lower barriers to developing large-scale AI models.
Similarly, the BigScience project brings together over 1,200 interdisciplinary researchers, emphasizing the importance of facilitating direct access to AI components across institutions and domains.
This open collaboration democratizes access to large-scale AI models, enabling researchers to fine-tune and adapt them for different languages and specific tasks, ultimately fostering a more diverse and representative AI landscape.
Open research and development also promote transparency and accountability in AI systems. For instance, the non-profit research organization LAION has released the open-source CLIP model, which plays a crucial role in identifying and addressing biases in AI applications. Open access to training data and model components allows researchers and the public to scrutinize the inner workings of AI systems and question misleading or erroneous claims.
The success of the AI Act depends on striking a balance between regulating and supporting open AI ecosystems. While openness and transparency are crucial, regulation must also mitigate risks, establish standards, and assign clear responsibilities for potential harms caused by AI systems.
As the EU establishes a regulatory framework for AI, embracing open-source and open science is essential to ensure that AI technology benefits all citizens.
By implementing the recommendations provided by organizations representing the interests of the open AI ecosystem, the AI Act can foster a collaborative, transparent, and innovative environment, positioning Europe as a leader in responsible AI technology development and application.
(Source: https://www.artificialintelligence-news.com/2023/07/26/ai-act-power-open-source-guiding-regulations/)
Alphabet Q1 paid subscribers reached 350M, fueled by YouTube & Google One. YouTube ad revenue slightly missed at $9.88B amid Premium growth. Total Q1 revenue hit $109.9B, beating forecasts.
Anthropic releases Claude connectors for Adobe, Blender, and Ableton, enabling AI workflow automation and app control. It also sponsors the Blender Development Foundation.
Apple introduces an App Store model splitting 12-month plans into monthly installments. On iOS 26.4+, it lowers upfront costs without reducing total price, adding billing tracking.
Google launches Aadhaar digital IDs in Wallet for India, enabling secure QR/NFC verification. A phased partner rollout targets key sectors, while passport-based credentials expand globally.
YouTube tests "Ask YouTube," an AI tool blending text and video for complex queries. Initially for US Premium users, it supports interactive follow-ups, expanding Google's AI search.
Google grants US DoD classified AI access, joining OpenAI and xAI in defense deals amid Anthropic's dispute. Contractual safeguards against surveillance and lethal weapons remain unenforceable.
Mistral AI launches Workflows, a Temporal-powered orchestration platform for secure, Python-driven enterprise AI automation featuring human-in-the-loop approvals.