Assuring Safe AI.
From ChatGPT to augmented reality, precision medicine, optimized supply chains, and personal AI assistants, the possibilities of artificial intelligence are endless.
The promise of AI is staggering. Improved decision making. Increased productivity. Enhanced customer service. Improved healthcare. Increased safety. Environmental sustainability.
Goldman Sachs predicts that AI could increase corporate profits by 30 percent in the next decade, and that generative AI alone could raise the global GDP by 7 percent.
But AI will also be used to violate personal privacy, spread disinformation, perpetuate social bias, and disrupt democratic processes.
The military kill chain will be driven at hypersonic velocity without a human in the loop. Cyberattack swarms will besiege critical infrastructure like 21st century locust plagues.
To forestall the deleterious consequences of AI, some are proposing the creation of regulatory agencies to review and approve computer algorithms prior to publication.
These proposals are na?ve. The number of algorithms published today is exceptionally large and growing exponentially. There is simply not enough money, not enough people, nor sufficient intellect to evaluate this enormous volume of mind-numbingly complex information. It is an impossible task.
The effects of regulation will be to hamstring technological advancement, or worse, to criminalize science—to the advantage of our economic competitors and military adversaries.
The most efficacious approach to control AI rapidly is to legislate, not regulate. A few simple, common-sense guidelines would go a long way toward mitigating risk.
It should be unlawful to:
- Publish algorithms that collect personally identifiable information without providing each person the right to be forgotten upon receipt of notice
- Publish algorithms that create or propagate hazards to public health
- Publish algorithms that are used to facilitate crimes
- Publish algorithms that are used to interfere with the democratic process
- Deploy high-consequence military AI applications, including but not limited to weapons systems, without senior civilian oversight
Straightforward legislation can leverage the well-established and rich fabric of the courts and criminal justice system to keep us safe.
Thomas M. Siebel
CEO, C3 AI
Strengthen businesses' financial performance through outsourcing services in F&A and Automation | FP&A | Business Analysis | Design processes to optimize use of Technology.
10 个月Thomas, thanks for sharing!
Founder & HoBD @Devxhub ?? Your Trusted Go-to Team Augmentation & Cost-Effective Software Development Partner. ??Hire Devxhub to make a difference.
10 个月Insightful perspective on managing AI's impact, Thomas M. Siebel. Emphasizing legislation over regulation to handle AI's rapid evolution is a critical approach to balancing innovation and ethical concerns. #AIethics #TechLeadership #InnovationGovernance
It's an epic journey, always.
1 å¹´Spot on. This allows the pace of innovation by the good actors to be as unrestricted as that of the bad actors. AI will clearly be used to legislate and must be free to keep pace.
English Teacher, English/ESL Department Chair
1 å¹´Thank you, Tom.
Better yet the legislative framework should become an integral part of the algorithms and AI models so that the AI can self regulate as it evolves and learns.