It’s almost war. Where is the line everyone could agree with?
https://strategicthought.ndu.edu.pk/index.php/site/article/view/142
INFORMATION WARFARE - THE INTERNATIONAL LEGAL REGIME EVOLUTION, CHALLENGES, ARCHITECTURE, AND THE PATH FORWARD
The use of information warfare is now increasingly fusing cyber intrusions, infrastructure compromises, disinformation, and synthetic media using artificial intelligence in ways that are blurring the lines between peace and armed conflict. International law is not lacking, but the rules are scattered between sovereignty and non-intervention, state responsibility, jus ad bellum, international humanitarian law (IHL), human rights law and platform regulation. The current debate on both topics is largely conducted in isolation: problems of technical cyber operations and cognitive information operations overlap in the issues of attribution, threshold, civilian harm, and state control of information flows. The article explores the UN Charter, ARSIWA, IHL, the Tallinn Manuals, UN cyber-governance processes, and selected state practices through the lens of qualitative doctrinal legal research and comparative case analysis. It looks at the cyber threat of Stuxnet, Salt Typhoon, and the India-Pakistan conflict of May 2025 as different types of information warfare. The analysis concludes that IHL cannot stand alone to regulate continuous activity that does not reach the threshold of armed conflict and that the treaty has a vital role to play in the regulation of information operations in situations where they are linked to armed conflict. It also suggests that the clash of different visions on digital sovereignty is an obstacle to norm formation. A reform agenda, based on the interpretation of the lex specialis, is proposed with a focus on attribution, critical civilian digital infrastructure, AI-generated synthetic media and accountability in the platform mediated information environment.
Mental Impressions as food for your AI?
https://escholarship.org/content/qt25z0n36h/qt25z0n36h.pdf
Generative AI and the Limits of Work Product Doctrine
The rapid integration of generative artificial intelligence into the legal practice has certainly created new efficiencies along with new uncertainties, particularly around confidentiality, discovery, and the work product doctrine. Given the recent advancements of AI, more specifically Large Language Models (LLMs), it seems the integration of AI into legal practice is inevitable.1 Due to large law firms already creating their own systems, 2 a natural consequence is that small firms, offices, or individuals that cannot afford to painstakingly create their own secure LLM3 will rely on OpenAI to even the field.4 Naturally, issues in regards to the work product doctrine will arise,5 and judges must decide when the work product doctrine should or should not bend. As judges delve into the issues of work product doctrine, it is important to ground these arguments with the explicit purpose of the work product doctrine: fairness.
In fact, courts are already beginning to confront whether these AI-assisted outputs are protected under traditional doctrines. In Tremblay v. OpenAI, Inc.,6 the district court faced this question directly when plaintiffs who used ChatGPT in presuit investigation sought to shield their prompts and outputs as attorney work product. Though a magistrate initially ordered disclosure, the district judge reversed, holding that prompts crafted by counsel reflected mental impressions and were therefore protected as opinion work product under Rule 26(b)(3).
