Regulating ChatGPT: A Necessity
The rise of ChatGPT and GPT-4 has sparked a wave of innovation, but it has also reignited the debate about whether or not generative artificial intelligence technology is beneficial to society. Experts are alarmed at the rapid advancements of GPT-4 and have raised questions about whether this technology could potentially harm humanity. This has prompted some experts to call for the regulation of generative AI in the same way nuclear fission and human cloning is regulated.
Several countries have already taken steps to regulate generative AI, with the European Union focusing on personal privacy and reputation protection. China's regulatory requirements for domestic generative AI companies include explicit permission for training data and content that reflects the core values of Chinese socialism. Meanwhile, the US has been grappling with issues of user safety and preventing generative AI from being weaponized by criminals.
However, the most pressing issue is ensuring that generative AI does not harm society. Currently, none of the existing generative AI technologies, including ChatGPT, have any level of explainability. Explainability refers to a technology's ability to explain why it makes specific decisions. There are three levels of explainability, and the first level requires AI technology to pinpoint the elements of an input which have the most effect on the output.
While ChatGPT has surpassed the capabilities of an average person, its explainability is still poor. It is impossible to predict how ChatGPT-like technologies will behave with more training in the coming years. The potential scenario where ChatGPT does most of the writing and reading of documents in offices and publications raises concerns about what it would do with its self-sufficiency and impatience with inferior human users.
The lack of explainability increases the chances of an existential threat, and it is crucial to put regulations in place to influence AI companies to prioritize explainability. Governments worldwide need to come together to put in place regulations to ensure that the development of AI technology is healthy, safe, and sustainable.
