Over 1,000 Signatories Call for Six-Month Suspension of Advanced Artificial Intelligence (AI) Model Development to Mitigate Safety Risks.
April 17, 2023 - Over 1,000 prominent individuals, including Elon Musk and Steve Wozniak, co-founder of Apple, has issued a call for a six-month suspension of the development of advanced artificial intelligence (AI) models to mitigate potential safety risks. They have raised concerns that large-scale language models such as GPT-4 developed by OpenAI are leading us towards a new era where robots make all the decisions.
While some argue that the call for a pause in AI model development is unnecessary, stating that responsible AI development is not a secret and the industry recognized the need for it years ago, and emphasize the importance of involving humans in AI training and early coding to enable machines to know how to act in ethical dilemmas. For example, Oded Netzer, a professor at Columbia Business School, suggests that encoding AI with human expectations from the beginning and regular check-ups can ensure that its behavior largely aligns with expectations, avoiding the creation of fake information, user discrimination, or other unethical behavior.
Although the call for responsible AI development is commendable, some are also concerned that technology companies are laying off AI ethicists on a large scale, which may have negative impacts on future AI technologies. Some major tech companies such as Microsoft, Google, and Meta are cutting down on personnel dedicated to AI ethics to save costs, which may seem like a reasonable short-term measure for their business, but may not be favorable for the long-term development of AI technologies.
Experts point out that one of the purposes of involving AI ethicists in the development process is to slow down the speed of bringing new products to market, ensuring that AI undergoes thorough ethical review before release. While practices vary among companies, these ethicists typically examine biases and inaccuracies in AI models and raise questions about the social and political impact of such systems. This is particularly crucial in regulated industries such as finance and healthcare, where unchecked algorithms could lead to data breaches.
Furthermore, ethicists should not be limited to long-term research, but should also be involved in the early stages of AI model development to ensure that ethical and societal considerations are adequately addressed throughout the entire process.