OpenAI, the developer of ChatGPT, has formed a new 'Safety and Security Committee' to review safety and ethical issues during the development of generative artificial intelligence (AI) technologies. This move comes in response to mounting criticism against OpenAI following the recent controversy over the voice imitation of famous actress Scarlett Johansson.
On the 28th (local time), OpenAI announced in a blog post that Brett Taylor, the chairman of the board, CEO Sam Altman, and board member Adam D'Angelo will lead the committee. The committee's first task is to evaluate OpenAI's processes and safety measures and to further develop them.
This decision was made shortly after OpenAI's recently launched voice service 'Sky' for 'GPT-4o' became embroiled in controversy for allegedly mimicking Johansson's voice. Following the release of Sky, online suspicions arose that Johansson's voice had been used without permission to train the AI, and Johansson herself stated that although OpenAI had requested to use her voice, she did not grant permission, which intensified the backlash. Ultimately, Johansson's side hinted at legal action, and OpenAI has since ceased using the voice.
As a result, criticism is growing that AI services are using various content, creative works, and even human voices without consent during the training process. There are also increasing calls for major companies, including OpenAI, to transparently disclose AI training data. The Wall Street Journal (WSJ) reported, "Generative AI chatbots rely on text, images, music, and videos across the internet, including copyrighted or paid materials," adding, "The rapid rise of AI is causing deep anxiety among those who have created content until now."
Previously, OpenAI disbanded its existing safety team, including reallocating part of the superalignment team that researched AI risks to other teams. Ilya Sutskever, OpenAI's chief scientist who led the team, also left OpenAI.
On the same day, OpenAI confirmed that it has begun training the next-generation AI model. OpenAI announced, "We have started training the next frontier model," and "We expect the outcome to lead to the next level of capabilities toward artificial general intelligence (AGI)." However, specific details such as the release timing of the next AI model, GPT-5, were not disclosed. CEO Altman and other OpenAI executives have consistently emphasized the goal of developing AGI that benefits humanity. AGI refers to AI with intelligence comparable to or exceeding that of humans.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
