Artificial intelligence accelerates progress in deep ways and transforms everything from our daily routines to the boundaries of scientific discovery and creativity. But as AI breakthroughs accelerate, how should we promote a balanced approach to the use of creative content in training AI models? This goes beyond legal technical conditions to affect the future of AI innovation and human creativity.
Each new technology for creating or transmitting knowledge and art, from the printing press to the Internet and Kabel -TV, has raised questions about how to create and share value. In the case of AI, developers can take a number of steps to support the creative industries and help build a flowering AI ecosystem that benefits everyone. What approaches makes sense for output from AI models, training these models and the new ways AI can create shared value?
Evaluation of AI -Outputs
Whether words are created with a pen, a typewriter or AI, or art is created with a brush, computer graphics or AI, the question is whether a new work violates a previous copyright. This judgment may be complex, depending on factors such as where the like the new work is for the older, the nature of the two works, and whether the new ones are competing in the original market. Tools such as output filters can help limit significantly similar output, even when the models themselves learn to make more nuanced assessments of these factors.
And origin information, such as watermarks or metadata, can reduce the risk of deception around the creator of particular material. For example, Google is paving the industry-leading Synthid tool and has joined the control committee for the coalition of content-making and authenticity (C2PA). These kinds of efforts can help consumers make informed assessments of the content they see.
Education of AI models responsible
While training of basic AI models on the available content on the open web is a transformational fair use under US copyright law, and many other countries have exceptions to text and data settings that similarly promote new uses of information, good practice can help build acceptance of new AI uses of existing content.
It is important to acquire content responsible and legally, e.g. By giving sites the opportunity to opt out of having content or information on their sites used for AI training. Existing industrial standards that control web crawling are an important way to achieve this. These standards are simple and scalable and are based on long-established machine-readable robot.txt protocols that are widely used across the Internet to check how their content is obtained by web crawlers. And now thousands of web publishers are also using the Google-Extended Protocol and similar AI-specific protocols offered by other companies. AI developers should remain open to developing these standards as the ecosystem progresses, and should take reasonable steps to avoid wrong to educate general purposes AI models in ways bypassing these standards or similar technical measures as payment walls.
When it comes to avoiding the use of individuals’ votes and equality, regulatory frameworks can build on existing “notice-and-removal” systems for copyright, including appropriate security measures to prevent abuse. New tools can also help creators to exploit AIS creative potential while letting them keep control of their voice and equality.
Sharing value, extended option
AI has the potential to benefit everyone, and cooperation between AI developers and content publishers can expand the market and generate new income for creative industries.
AI developers seek to share the value of output by sending related traffic to content providers. And the ecosystem works together to find new ways to create value of new AI applications. For example, there may be opportunities for commercial partnerships when AI services “Earth” answers to facts from websites.
AI developers and content publishers also work together on new content agreements for the use of specialized or non-public data for training purposes. AI developers are increasingly learning to assess the utility of individual content for different AI applications. For our part, Google has already entered into agreements with several publishers on broad dating rights, and we continue to explore new opportunities.
AI developers are actively working with media and creative industries to design new generative AI tools that add value to these industries. E.g. Helps an AI tool for journalists, journalists searching through text, audio, image and video files to see patterns in data, identify new angles, or find an offer in a video or audio file.
AI is a common opportunity with the potential to expand the world of science, trade and creativity. We are obliged to cooperate with all stakeholders in the ecosystem to create a common framework where both the rights and innovation of the creator flourish.