127 disclosure and general teaching. Israel recognizes the “essence of the invention,” and if that essence has become public, it must follow that the invention is no longer novel. This is especially problematic because the novelty inquiry is unforgiving: there is no general grace period for the inventor’s own disclosures. A single leakage event before filing can be fatal. United States: Prior Art and Grace Periods Under § 102 US law defines prior art broadly, including patents, printed publications, public use, on-sale activity, and other public disclosures before the effective filing date. The America Invents Act introduced a limited one-year grace period for disclosures made by the inventor or derived from the inventor. In a best-case scenario, an inventor who disclosed to AI might argue that subsequent dissemination was derived from the inventor and therefore falls within the grace period. However, that argument is evidentiary and litigationheavy. Who can prove what the inventor disclosed to AI? Who can prove that later AI outputs were “derived” from that disclosure? Can the inventor obtain logs? Will a provider produce them? And even if derived, the grace period does not help outside the United States. A global filing strategy cannot be built on a fragile domestic safety net. Novelty and Prior Art: Practical Risk Scenarios To make these issues concrete, consider three recurring scenarios in practice. Scenario A: “Is this idea patentable?” An inventor pastes a detailed invention disclosure and asks an AI system to assess patentability and draft claims. Even if the AI response is helpful, the act of disclosure creates a risk that the core idea enters the model’s generalized behavior or a provider’s data pipeline. If a similar prompt later yields a similar claim set for another user, novelty attacks may follow, especially in Israel and in Europe. Scenario B: “Help me solve this technical bottleneck.” An engineer describes a specific technical problem and provides experimental constraints. The AI proposes a solution. If the solution is later echoed elsewhere, an opponent may argue that the solution was publicly available at the relevant date. Even if the AI’s solution is obvious in hindsight, the presence of AI-mediated dissemination complicates the novelty story and can undermine perceived inventiveness. Scenario C: “Translate and improve my draft.” A practitioner provides a draft application containing proprietary details and asks the AI to improve clarity, restructure embodiments, and propose dependent claims. This blends mechanical editing with substantive invention disclosure. It also risks introducing AI-generated language that is not fully understood or verified, which can later become a Sec. 4/§112/Art.83 vulnerability. These scenarios illustrate a practical rule: the closer AI use gets to inventive substance and strategic claim architecture, the higher the legal risk. Inventive Step / Non-Obviousness: The “Skilled Person” Meets AI Even when novelty survives, inventive step (IL Sec. 5 and EPC Art. 56) and non-obviousness (35 U.S.C. § 103) may be affected by widespread AI use. These doctrines ask whether the invention would have been obvious to a person skilled in the art at the relevant date. Europe: Problem–Solution Approach Under the EPC EPO practice typically applies the problem–solution approach: identify the closest prior art, define the objective technical problem, and ask whether the claimed solution would have been obvious. AI tools can influence each step. They can rapidly identify the candidate closest prior art, propose problem formulations, and suggest “routine” modifications. As AI becomes a common tool in the field, opponents may argue that certain combinations are standard because AI makes them easy to find. A subtle but important point is that the EPO’s skilled person is not omniscient, but is assumed to have access to the common general knowledge and to routine experimentation. If AI-assisted searching and reasoning become routine, the baseline may shift in practice. Practitioners should anticipate arguments that “any skilled person would have found this” because an AI tool can do so quickly. While that argument should not be accepted uncritically, it may become persuasive in oppositions where fact patterns matter. ISRAEL — INTELLECTUAL PROPERTY
RkJQdWJsaXNoZXIy MjgzNzA=