The conduct of the actions by the providers and prospective providers in the AI regulatory sandbox should remain under the responsibility of the participants themselves, including as regards complying with any relevant legal provisions applicable to them. The establishment of an AI regulatory sandbox should not serve as a basis for limiting or excluding the liability of participants under this Regulation or any other applicable Union or national law, including as regards any damage to third parties suffered as a result of activities carried out in the AI regulatory sandbox. The participants in the AI regulatory sandbox should be required to appropriately document the results of the testing and experimentation processes, including as regards their assessment of potential risks and other matters that may be relevant for decision-making regarding the placing on the market or putting into service of AI systems.