The Gladstone Report's Call to Action: Beyond the EU AI Act Towards Global AI Safety
Navigating the AI Maze: The Quest for Global Regulation
On Wednesday, the European Parliament took a significant step forward by approving the Artificial Intelligence Act. This legislation has stringent requirements for high-risk AI systems in critical sectors, including risk assessment, transparency, and human oversight, and empowers citizens to seek explanations for AI-driven decisions. Additionally, it imposes transparency and compliance obligations on general-purpose AI systems, such as the clear labelling of manipulated content like deepfakes.
Social media’s company X/Twitter recent refusal to remove a deepfake audio clip misrepresenting UK opposition leader Keir Starmer underscores the ongoing battle against electoral interference. Despite the clip's rapid proliferation and Labour's requesting its removal, X has hesitated, citing policy ambiguities, which starkly contrasts with the proactive stance taken by other platforms like Facebook, YouTube, Snapchat, and TikTok. This incident not only highlights the regulatory challenges of synthetic media, but also amplifies concerns about the misuse of artificial intelligence in politics, and the readiness of both political entities and social media firms to counter disinformation effectively.
The Gladstone report – a comprehensive analysis commissioned by the US State Department's Bureau of International Security & Nonproliferation – underscored the urgency for further action beyond the EU AI Act. It highlights the national security risks presented by the rapid advancement of frontier AI and AGI (Artificial General Intelligence), comparing its potential for global destabilisation to that of nuclear weapons. The report aligns with Control AI’s propositions, notably calling for heightened accountability across the AI supply chain and specifying that different sectors, from hardware designers to data centres, adhere to detailed compliance instructions.
Significantly, the report also proposes thresholds for AI development, suggesting prohibitions on advancements beyond certain points – albeit with regulatory flexibility to adjust these limits based on AI evolution. It also emphasises the need for comprehensive emergency plans across the supply chain to ensure swift compliance with government-ordered shutdowns in the event of an AI-induced crisis. As we mark this global milestone in AI regulation with the EU AI Act, the Gladstone report reminds us of all the work that remains.
If you want to delve deeper into the challenges of AI governance, the regulation of synthetic media, and the global security implications of AI advancements, join us on Discord at https://discord.gg/32EXTWNEfe. Here, we can collaborate, share insights, and contribute to shaping the future of AI in a manner that safeguards our security, democratic values, and fosters responsible innovation.