In an initiative to promote ethical AI practices, seven leading AI tech firms, among them Google, OpenAI, and Meta, have allied with the US government under the leadership of President Joe Biden. This collaboration aims to tackle potential risks that arise from AI technology by implementing new standards and safeguards for secure and ethical AI development.
As per reports by IANS, Amazon, Anthropic, Meta, Google, Inflection, and OpenAI are part of this groundbreaking agreement. The proposed actions include performing security audits on AI systems and openly sharing the results. This step towards transparency plays a pivotal role in establishing accountability and fostering public trust.
During a discussion held at the White House, President Biden underlined the necessity for ethical AI practices. He recognized the tremendous influence AI could have globally and emphasized the responsibility that the entities involved have in ensuring this technology is developed in a safe and responsible manner.
In a statement, Meta’s president of global affairs, Nick Clegg, said, “AI should be beneficial for all of society, and for that to occur, we need to responsibly and safely build and deploy these potent new technologies.”
He further stated, “As we devise new AI models, it’s crucial that tech firms are transparent about how these systems operate and foster collaboration across sectors, including industry, government, academia, and civil society.”
One of the measures includes integrating watermarks into AI-generated content, thereby facilitating easier recognition for users. Regularly releasing reports about AI capabilities and restrictions will further contribute to a transparent AI environment.
The firms will also conduct research addressing the various risks connected with AI, like bias, discrimination, and invasion of privacy, ensuring AI technologies are developed in an ethical and responsible manner.
OpenAI stated that the watermarking agreements would necessitate companies to create tools or APIs to ascertain if specific content was generated through their AI systems. Google had earlier committed to implementing similar disclosures this year.
Furthermore, Meta has recently declared its decision to open-source its large language model, Llama 2, thereby making it accessible to researchers at no cost.