Guaranteed Safe AI Systems: A Solution for the Future of AI Safety?
Joerg Hiller Jun 26, 2024 02:40
Exploring the potential of guaranteed safe AI systems in ensuring the safety and reliability of artificial general intelligence (AGI).
In a recent discussion, Dr. Ben Goertzel, CEO of SingularityNET (AGIX), and Dr. Steve Omohundro, Founder and CEO of Beneficial AI Research, explored the critical issue of artificial general intelligence (AGI) safety. The conversation delved into the necessity of provable AI safety and the implementation of formal methods to ensure that AGI operates reliably and predictably, according to SingularityNET.
Insights from Decades of Experience
Dr. Steve Omohundro’s extensive background in AI, which began in the early 1980s, positions him as a leading voice in AI safety. He emphasized the importance of formal verification through mathematical proofs to ensure AI systems operate predictably and securely. The discussion highlighted advancements in automated theorem proving, such as Meta’s HyperTree Proof Search (HTPS), which have made significant progress in verifying AI actions.
Despite these advancements, applying automated theorem proving to AGI safety remains a complex challenge. The conversation also touched on various approaches to improving AI’s reliability and security, including provable contracts, secure infrastructure, cybersecurity, blockchain, and measures to prevent rogue AGI behavior.
Potential Risks and Solutions
Dr. Omohundro discussed the development of a programming language called Saver, designed to facilitate parallel programming and minimize bugs through formal verification. He stressed the fundamental need for safe AI actions as these systems become more integrated into society. The concept of “provable contracts” emerged as a key solution, aiming to restrict dangerous actions unless specific safety guidelines are met, thereby preventing rogue AGIs from performing harmful activities.
Building a Global Infrastructure for AI Safety
Creating a global infrastructure for provably safe AGI is a monumental task that requires significant resources and global coordination. Dr. Omohundro suggested that rapid advancements in AI theorem proving could make verification processes more efficient, potentially making secure infrastructure both feasible and cost-effective. He argued that as AI technology advances, building secure systems could become cheaper than maintaining insecure ones due to fewer bugs and errors.
However, Ben Goertzel expressed concerns about the practical challenges of implementing such an infrastructure, especially within a decentralized tech ecosystem. They discussed the need for custom hardware optimized for formal verification and the potential role of AGI in refactoring existing systems to enhance security. The idea of AGI-driven cybersecurity battles also came up, highlighting the dynamic and evolving nature of these technologies.
Addressing Practical Challenges and Ethical Considerations
The discussion also addressed the significant investment required to achieve provably safe AGI. Ben Goertzel noted that such initiatives would need substantial funding, potentially in the hundreds of billions of dollars, to develop the necessary hardware and software infrastructure. Dr. Omohundro highlighted the progress in AI theorem proving as a positive sign, suggesting that with further advancements, the financial and technical barriers could be overcome.
Ethical considerations were also a critical part of the dialogue. Ben Goertzel raised concerns about large corporations pushing towards AGI for profit, potentially at the expense of safety. He emphasized the need for a balanced approach that combines innovation with robust safety measures. Both experts agreed that while corporations are driven by profit, they also have a vested interest in ensuring that their technologies are safe and reliable.
The Role of Global Cooperation
Global cooperation emerged as a key theme in developing beneficial AGI. Steve Omohundro and Ben Goertzel acknowledged that building a secure AI infrastructure requires collaboration across nations and industries. They discussed the potential for international agreements and standards to ensure that AGI development is conducted safely and ethically.
This insightful discussion underscores the complexities and opportunities in ensuring a secure and beneficial future for AI. By fostering more cooperation in the field, advancing a safe and predictable path for AGI development, and addressing ethical considerations, the vision of a safe and harmonious AI-driven future is within reach.
Image source: Shutterstock