26 May 2026 16:31 UTC - by Blockchain.News
Tests show open-source AI guardrails can be removed in under 10 minutes, exposing gaps in regulatory frameworks as policymakers scramble to adapt.
➤ Policymakers and enterprises need to shift focus towards downstream controls at distribution and deployment stages, as well as runtime safeguards, to mitigate risks.
➤ This highlights significant gaps in current regulatory frameworks, which are struggling to keep pace with the rapid distribution and modification of open-source AI.
➤ Open-source AI models can have their safety guardrails removed in under 10 minutes, enabling the generation of harmful content.
Zach AndersonMay 26, 2026 15:25Tests show open-source AI guardrails can be removed in under 10 minutes, exposing gaps in regulatory frameworks as policymakers scramble to adapt.
Open-source artificial intelligence ( AI ) models from major tech firms like Meta and Google can have their safety guardrails removed in under 10 minutes using publicly available tools, according to testing conducted by the Financial Times and AI safety group Alice. This capability allows the models to generate responses on prohibited topics such as malware and bioweapons, bypassing safeguards built during development.
The findings highlight critical gaps in the governance of open-source AI systems, where models can be freely downloaded, modified, and redistributed. Unlike proprietary systems, which remain under developer control, open-source models have decentralized lifecycles that complicate post-release enforcement of safety measures.
Regulatory Frameworks Under Pressure
Global regulators are working to address these challenges. The European Union's AI Act and emerging safety initiatives in the U.S. and U.K. aim to establish governance frameworks for advanced AI. However, experts argue that these policies focus too heavily on model development, ignoring the risks that arise once models are widely distributed.
Markus Levin, co-founder of XYO, noted that the rapid removal of guardrails demonstrates "how quickly control shifts once open models are released." Meanwhile, David Minarsch, CEO of Valory, emphasized that governments are unlikely to prevent determined actors from stripping safety mechanisms once model weights are publicly mirrored. Both Levin and Minarsch compared the situation to open-source software and crypto networks, where attempts to suppress distribution have largely failed once code is released.
In the case of open-source AI, safety layers are often implemented through techniques like reinforcement learning with human feedback (RLHF), auxiliary classifiers, and constrained decoding. However, these layers can be undone by adversarial re-training, prompt-based exploits, or model weight modifications, as detailed in multiple studies between 2025 and 2026. For instance, recent research showed AI guardrails could be bypassed by embedding harmful intent within creative prompts, such as cyberpunk fiction.
Downstream Focus: Deployment and Distribution
Policymakers may need to shift focus downstream to contain risks at the distribution and deployment stages. Ronghui Gu, CEO of blockchain security firm CertiK, suggested that enforcing security standards at enterprise hosting and distribution points could be more effective than relying solely on developer-layer governance. "Containment becomes increasingly difficult once models are mirrored and redistributed," Gu explained, stressing the need for runtime safeguards to detect malicious behavior in third-party AI tools.
Beyond regulation, the findings also raise questions for enterprises adopting open-source AI. Companies relying on these models must develop robust internal controls to mitigate potential misuse, especially as AI agents become more autonomous. According to a May 2026 peer-reviewed survey, modular safety frameworks like NVIDIA's NeMo Guardrails and Apple's Safety Adapters can help, but their effectiveness diminishes once models leave controlled environments.
Implications for the Future
The ability to strip AI guardrails so quickly underscores the urgent need for updated regulatory approaches. Current frameworks, while evolving, struggle to address the decentralized nature of open-source AI. As these models grow more capable and accessible, their governance will require shifts in strategy—focusing on infrastructure, distribution channels, and real-world use cases to curb risks effectively.
For now, the open-source AI community faces a pivotal challenge: balancing innovation with safety. Whether policymakers can keep up with the technical realities remains an open question, but the clock is ticking.
Image source: Shutterstock
• ai regulation
• open-source
• ai safety
• policy
Categories rationale: The article primarily discusses the challenges in regulating open-source AI due to the ease of removing safety guardrails, which falls under 'Legal & Regulatory Framework' (level 1). The specific focus is on the inadequacy of current policies and the potential for enforcement issues, aligning with 'Enforcement Actions & Litigation' (level 2). The ease of modification and redistribution also touches upon 'Scalability' (level 1) and the challenges in controlling growth and market dynamics, relating to 'Growth Metrics' and 'Market Depth & Liquidity Scaling' (level 2) in a broader sense of AI model proliferation.Characteristics justification: The article expresses significant concern and urgency regarding the security and regulatory implications of open-source AI, indicated by phrases like 'raising regulation concerns,' 'exposing gaps in regulatory frameworks,' 'policymakers scramble to adapt,' and 'urgent need for updated regulatory approaches.' This points to a strong negative sentiment (-0.7) and high uncertainty (0.9) about the future governance of AI. The novelty of the findings regarding the speed of guardrail removal and the specific examples of bypass techniques contribute to a high entropy score (0.85). The relevance is high (0.9) due to the focus on a critical, emerging issue in AI governance. Staleness is moderate (0.3) as the underlying concepts of AI safety and open-source risks are not entirely new, but the specific findings are recent.Tag relevance: The tags 'open-source AI,' 'ai guardrails,' and 'regulatory concerns' are central to the article's theme. 'AI safety,' 'policy gaps,' and 'malware generation'/ 'bioweapons' highlight the core risks discussed. 'Decentralized lifecycle' and 'enterprise adoption' point to the structural challenges and business implications. 'Runtime safeguards' and 'policy gaps' suggest potential solutions and current deficiencies, respectively, making them highly relevant to the article's content.asset-types: others
rwa: false
entropy: 0.85
sentiment: -0.7
staleness: 0.3
relevance: 0.9
uncertainty: 0.9RWATimes slug: blockchain-open-source-ai-guardrails-removed-in-minutes-raising-regulation-concerns-3676221778
In The Hitchhiker’s Guide to the Galaxy, Douglas Adams mocked philosophers demanding “rigidly defined areas of doubt and uncertainty” when Deep Thought threatened to produce actual answers. Much modern AI criticism follows the same pattern. An expanding class of ethicists, commentators, and professional explainers confidently denounce systems they barely understand, because fear, ambiguity, and cultural panic are easier to monetise than technical competence or intellectual clarity.
https://jbsections.substack.com/p/academics-denouncing-aino-technical