Generative AI in Media & Journalism: Think Big, But Read the Small Print First
Terms of Use — typically hidden away at the bottom of a page, barely readable with the naked eye and notorious for their use of legalese– are where the use of generative AI systems is currently being governed. The Terms of Use of generative AI providers such as Open AI, Midjourney, Stable Diffusion, and others lay the ground rules for the responsible use of generative AI, instruct downstream users how to mitigate potential safety risks and risks to the realisation of fundamental rights, and create binding legal obligations. Recent legal initiatives such as the draft AI Act are likely to increase the importance but also the weight of the Terms of Use and Usage policies because it is here that providers of generative AI instruct users how (or how not) to use the technology and distribute responsibility. At the same time, the AI Act is also likely to trigger closer scrutiny and trigger a debate on the fairness of some of those terms. In the following post, we will critically and thematically review the Terms of Use and usage policies of five generative AI providers and recommend a number of terms to the attention of media organisations that are planning to use the models.[1] We included proprietary and non-proprietary, open systems and platforms in the selection, including Open AI, Midjourney, Anthropic, Hugging Face, and StabilityAI.Before starting to use or integrate a particular generative AI model media organisations would do well to carefully read the Terms of Use. Different providers make their models available under different conditions.
0 Comments