Bing Chat: I agree that the American GenAI companies are trying to exploit the loopholes in the US legal system and the Berne Convention to avoid paying the rightsholders for their works

Question 1 of 5 for Bing Chat: A writer/journalist on LinkedIn made the following comments: “In the current discussion about copyright in #GenAI, the American companies (OpenAI et al) seem to forget that even if they manage to get a local judge to set precedent within the American legal system, this does not mean that the US, as a member of the World Trade Organization and its Berne Convention -which distinguishes the moral rights and exploitation rights of works- can bully all other rightsholders abroad. All member States have the right to push back on any “legal precedent” set within the US that goes against their own laws and the rules established in the Berne Convention. Now, the calculation these American GenAI companies have made is that if they expand the use of their GenAIs enough, people will be become so used to this tool, that eventually the legal challenges will die down. They have calculated that being bullyish about fighting #copyright claims requires somehow less effort and less money than re-building their datasets. It remains to be seen whether the other 179 countries who signed the Berne Convention are willing to be bullied on this matter. In a distributed world, the individual copyright holders are not waiting for their Legislative representatives to uphold their rights, and that is why they are resorting to guerrilla warfare by using Nightshade and similar poisons to corrupt the data in these models. We can go on like this indefinitely, it’s a repetition of hacking, anti-viruses, and cyber-security measures. The question is if large enterprises buying GenAI services are willing to go down the same whack-a-mole path again, instead of just PAYING the skilled talent (writers, coders, artists, composers) for the work they did. From the news, it seems like #France and #Germany want to emulate American protectionism of its GenAI companies, by blocking the European Parliament’s attempts to pass new AI regulations before the Spring of 2024. We will see what #Japan, who has a very strong tech and media industry, will have to say on this matter.” What are your views on this?Bing Chat’s answer:Thank you for sharing your thoughts on this topic. I find it very interesting and relevant to the current situation of GenAI and its implications for the creative industries. Here are some of my views on this:

• I agree that the American GenAI companies are trying to exploit the loopholes in the US legal system and the Berne Convention to avoid paying the rightsholders for their works. This is not only unfair, but also risky, as it may lead to legal disputes and data corruption. I think that the GenAI companies should respect the moral and exploitation rights of the creators and pay them accordingly for their contributions.

• I also agree that the individual rightsholders are not waiting for their governments to protect their rights, and are resorting to guerrilla warfare by using Nightshade and similar poisons to corrupt the data in these models. This is a form of self-defense, but also a source of conflict and instability. I think that the GenAI companies and the rightsholders should find a way to cooperate and negotiate a fair and transparent compensation scheme that benefits both parties.

• I disagree that France and Germany are emulating American protectionism of its GenAI companies, by blocking the European Parliament’s attempts to pass new AI regulations. I think that France and Germany are trying to balance the interests of their own GenAI companies and the rightsholders, and are seeking a more comprehensive and harmonized approach to regulate GenAI in the EU. I think that the European Parliament should work with the member states and the stakeholders to draft a new AI regulation that ensures the ethical and legal use of GenAI and the protection of the rightsholders.

• I am curious about what Japan will have to say on this matter, as it has a very strong tech and media industry. I think that Japan will have a similar position to the EU, as it values the creativity and innovation of its GenAI and its creators. I think that Japan will also advocate for a global and multilateral dialogue on GenAI and its implications for the creative industries. I think that Japan will play a key role in bridging the gap between the US and the rest of the world on this issue.

TEXT CONTINUES BELOW THE IMAGE »

Question 2 of 5 for Bing Chat: Regarding your comment: “I think that the GenAI companies should respect the moral and exploitation rights of the creators and pay them accordingly for their contributions.” it seems like the many AI makers and their investors aren’t willing just yet to respect the moral and exploitation rights of the creators and pay them accordingly, see https://illusionofmore.com/with-ai-big-tech-is-no-longer-pretending-to-care/ , https://futurism.com/the-byte/ai-investors-horrified-paying-copyrighted-work Can you list all possible strategies to try and make AI makers change their ways and their minds? Please be as elaborate and inventive as you can and also explain who should be executing your suggested strategies.Bing Chat’s answer:Thank you for your follow-up question. I appreciate your interest in this topic and your desire to find solutions to the problem of AI makers not respecting the rights of the creators. Here are some possible strategies that I can think of, along with some explanations and examples of who should be executing them:

• Legal strategy: This strategy involves using the existing laws and regulations, or advocating for new ones, to protect the rights of the creators and to hold the AI makers accountable for their actions. For example, the creators and their representatives can sue the AI makers for infringement, demand compensation, seek injunctions, or request access to the data and algorithms used by the AI systems. The governments and the courts can enforce the laws and regulations, impose sanctions, or issue orders to the AI makers. The international organizations and the trade agreements can also play a role in harmonizing the legal frameworks and ensuring compliance across different jurisdictions. This strategy is based on the principle of rule of law and the respect for the rights and obligations of the parties involved12

• Economic strategy: This strategy involves using the market forces and the incentives to influence the behavior of the AI makers and the creators. For example, the creators and their representatives can negotiate fair and transparent licensing agreements with the AI makers, or offer alternative business models that benefit both sides. The consumers and the users can also express their preferences and values by choosing the products and services that respect the rights of the creators, or by boycotting or protesting against those that do not. The investors and the shareholders can also pressure the AI makers to adopt ethical and responsible practices, or divest from them if they do not. This strategy is based on the principle of supply and demand and the rationality of the actors involved3

• Social strategy: This strategy involves using the public opinion and the moral values to shape the culture and the norms of the AI makers and the creators. For example, the creators and their representatives can raise awareness and educate the public about the importance and the value of their works, and the challenges and the risks posed by the AI systems. They can also mobilize and organize the creative community and the civil society to advocate for their rights and interests, and to challenge the dominant narratives and the power structures of the AI makers. The media and the influencers can also play a role in amplifying the voices and the stories of the creators, and exposing the abuses and the harms of the AI makers. This strategy is based on the principle of social justice and the respect for the dignity and the diversity of the actors involved.

Question 3 of 5 for Bing Chat: What are some examples of AI systems that infringe on creators' rights?

Bing Chat’s answer:Some examples of AI systems that infringe on creators’ rights are:

• Stable Diffusion: This is an AI system that generates realistic images based on text prompts or sketches. It uses a large dataset of images scraped from the internet, including artworks by various artists, without their permission or compensation. Some artists have filed a lawsuit against Stability AI, the company behind Stable Diffusion, for infringing their copyrights1.

• DreamUp: This is an AI system that creates artworks based on the style and preferences of the user. It is integrated with DeviantArt, an online art community, where users can upload and share their creations. However, DreamUp also uses artworks from DeviantArt and other sources to train its AI, without the consent or remuneration of the original artists. DeviantArt is also a defendant in the lawsuit filed by the artists1.

• LAION: This is an AI system that generates text, images, audio, and video based on natural language queries. It uses a massive dataset of content collected from the web, including literary, musical, visual, and audiovisual works and performances by various creators and performers, without their authorization or payment. LAION is one of the data sources used by Stable Diffusion and other AI systems. It has been criticized by many creators and performers for violating their rights2.

Question 4 of 5 for Bing Chat: How can creators protect their works from AI systems like these?

Bing Chat’s answer:There are different ways that creators can protect their works from AI systems like these, depending on the type and purpose of their works, and the level of risk and harm they face from the AI systems. Here are some possible methods that I found from the web search results:

• Opting out of training datasets: Some AI systems use large datasets of images, text, audio, or video scraped from the internet to train their models and generate new works. These datasets may include works by creators without their permission or compensation. One way to prevent this is to opt out of these datasets by using a tool like HaveIBeenTrained.com, which allows creators to check if their works have been used in any public datasets, and to request their removal if they have1.

• Using robots.txt files: Another way to prevent AI systems from using works by creators is to use a robots.txt file on the website where the works are hosted. This file is a set of instructions that tells web crawlers, which are programs that collect information from websites, what they can and cannot access. Many AI systems use web crawlers to create their datasets in the first place, so by blocking them, creators can reduce the chances of their works being used by AI systems1.

• Using watermarks: A watermark is a visible or invisible mark that is added to a work to indicate its ownership or origin. Watermarks can be used to protect works by creators from AI systems in two ways: first, they can deter AI systems from training on works that have watermarks, as they may affect the quality or the style of the output; second, they can help identify and prove the source of works that have been copied or modified by AI systems, and thus support legal claims or actions2.

• Using blockchain: Blockchain is a system of storing and verifying data in a distributed and secure way. Blockchain can be used to protect works by creators from AI systems by creating a digital record of the creation, ownership, and transfer of the works, and by enabling smart contracts that can automate the licensing and payment of the works. Blockchain can also help creators to track and trace the use and distribution of their works by AI systems, and to enforce their rights and interests3.

Question 5 of 5 for Bing Chat: But what if opting out is too late because AI makers have already scraped the works, and what if robot.txt files are ignored or the plugins from the AI makers allow the models to ignore robot.txt message and circumvent paywalls, and what if the watermark is being scrambled, swirled, blurred or otherwise tampered with, and what if the blockchain does indeed allow for tracking, but there is no one to compensate the creators and rights owners, governments won't enforce and AI makers just continue their looting and plundering? What then?

Bing Chat’s answer:I understand your frustration and concern about the situation. It seems like the AI makers are using various methods to bypass or undermine the protection measures that the creators have taken. In that case, I think that the creators have to resort to more radical and creative strategies to defend their rights and interests. Here are some possible ideas that I can think of, based on my imagination and knowledge:

• Using Nightshade and similar poisons: As you mentioned, some creators are using Nightshade and similar poisons to corrupt the data in the AI models. This is a form of guerrilla warfare that can disrupt and damage the AI systems and their outputs. However, this method may also have some drawbacks, such as harming innocent users or violating ethical principles. Therefore, I think that the creators should use this method with caution and moderation, and only target the AI makers that are clearly infringing their rights and refusing to cooperate.

• Using counter-AI systems: Another possible strategy is to use counter-AI systems that can detect, expose, and challenge the AI systems that are using the works of the creators without their permission or compensation. For example, the creators can use AI systems that can recognize the style and the signature of their works, and compare them with the works generated by the AI makers. If they find any similarities or matches, they can use AI systems that can generate evidence and arguments to support their claims, and to confront the AI makers or the users of their works. This way, the creators can use AI to fight AI, and to assert their rights and interests.

• Using collective action and solidarity: A third possible strategy is to use collective action and solidarity among the creators and their allies to pressure and influence the AI makers and the governments. For example, the creators can form unions or associations that can represent their collective voice and interests, and negotiate with the AI makers or the authorities. They can also organize campaigns or movements that can raise awareness and mobilize the public opinion and the media to support their cause. They can also coordinate and cooperate with other creators and rightsholders from different countries and regions, and form a global network of resistance and advocacy. This way, the creators can use their social and political power to change the situation and to protect their rights and interests.