AI content will be watermarked
OpenAI, Google, Meta, Microsoft committed to the US government to apply measures such as marking AI-generated content to ensure safety.
"This commitment is a promising step, but there is a lot more work to be done together," US President Joe Biden said at an event at the White House on July 21. "We must be discerning and vigilant about emerging technology threats to American democracy."
The event was attended by representatives of big companies in artificial intelligence such as OpenAI, Google, Meta, Anthropic, Inflection, Amazon, Microsoft. Together, they will develop an automatic watermarking tool for all formats from text and audio to photos and videos to make it easier for users to distinguish between human- and AI-generated content . The parties have not yet discussed how the watermark will be displayed.
In addition, the companies also promised to thoroughly test AI systems before release, share information on how to reduce risks, and will invest heavily in cybersecurity. They also claim to focus on protecting privacy and ensuring that artificial intelligence does not discriminate against users. Other commitments include developing AI-powered solutions for science such as medical research or climate change mitigation.
AI models are evolving. Photo: 3AW
The move is seen as a victory for the Biden administration's efforts to control generative AI , a technology that is exploding both in terms of investment and popularity with users.
Since the birth of AI like ChatGPT, Bard appeared and caused fever, legislators around the world have begun to consider measures to reduce the risk of this new technology to national security as well as the national economy.
Mr. Biden said he was working on an executive order on AI after receiving representatives of 7 technology companies. The US is considered to be lagging behind Europe in making AI regulations. In June, European lawmakers released a draft set of rules that would require systems like ChatGPT to distinguish AI-generated content from real content and ensure protections against malicious content.