Responsible AI Development
On March 22, 2023, the Future of Life Institute[^future-of-life-institute] published an open letter proposing "all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4." They argue that such systems pose profound risks to society and humanity and that they should be developed only once we are confident that their effects will be positive and their risks will be manageable.
[^future-of-life-institute]: The mission of the Future of Life Institute, established in 2015, is to steer transformative technology towards benefitting life and away from extreme large-scale risks.
The arguments of the letter are vague and generic, and not specific to AI. For example, the fear that "machines flood our information channels with propaganda and untruth", or that all the jobs get automated, and that those decisions are delegated to unelected tech leaders. It's not clear how a ban on training new Large Language Models addresses these problems.
The call to pause the training of AI systems more powerful than GPT-4 has received mixed reactions from the AI community. While some researchers and experts share concerns about the potential risks of developing such systems, others argue that a ban on research and development is not practical or effective. Also, one should note that many of the people who signed the Future of Life Institute's letter are themselves involved in AI. For example, Elon Musk signs the letter but at the same time is investing in AI.
Serious AI researchers that worry about AI are not worried about sentient AIs or about jobs lost to automation. They are worried that a sufficiently capable Large Language Model gets "out of control" by misinterpreting or trying to maximize its programmed goals in ways unthought by its creators and may endanger humanity by doing so.
These worries are partially fueled by the now-more-plausible hypothesis that COVID-19 was engineered in a laboratory in China and released by accident into the world. The AI system would be, in this case, like the COVID-19 virus. Developed by a government or company behind closed doors, or used by the industrial military complex to create warfare. Once released, it could be very difficult or impossible to control and be harmful to humanity.
Ultimately, the attitude one should adopt regarding the Future of Life Institute's letter should be one of critical thinking and open-mindedness. While it is important to take seriously the potential risks associated with AI development, we must also recognize the benefits that Artificial Intelligence can bring to society and work towards developing ethical and responsible frameworks for AI research and deployment.
Rather than calling for a blanket ban on Artificial Intelligence research and development, we should focus on addressing specific risks and challenges associated with AI, such as bias, transparency, and accountability. This will require collaboration between researchers and other stakeholders to ensure that AI is developed and deployed in a way that benefits humanity and aligns with our values and ethics.