Autonomous vehicles are now the future of automobile industry. Human drivers can be completely taken out of the loop through the implementation of safe and intelligent autonomous vehicles. Although we can say that HW and SW development continues to play a large role in the automotive industry, test and validation of these systems is a must. The ability to test these vehicles thoroughly and efficiently will ensure their proper and flawless operation. When a large number of people with heterogeneous knowledge and skills try to develop an autonomous vehicle together, it is important to use a sensible engineering process. State of the art techniques for such development include Waterfall, Agile & V-model, where test & validation (T&V) process is an integral part of such a development cycle. This paper will propose a new methodology using machine learning & deep neural network (AI-core) for lab & real-world T&V for ADAS (Advanced driver assistance system) and autonomous vehicles. The methodology will initially connect T&V of individual systems in each level of development and that of complete system efficiently, by using the proposed phase methodology, in which autonomous driving functions are grouped under categories, special T&V processes are carried on simulation as well as in HIL systems. The complete transition towards AI in the field of T&V will be a sequence of steps. Initially the AI-core is fed with available test scenarios, boundary conditions for the test cases and scenarios, and examples, the AI-core will conduct virtual tests on simulation environment using available test scenarios and further generates new test cases and scenarios for efficient and precise tests. These test cases and scenarios are meant to cover all available cases and concentrate on the area where bugs or failures occur. The complete surrounding environment in the simulation is also controlled by the AI-core which means that the system can attain endless/all-possible combinations of the surrounding environment which is necessary. Results of the tests are sorted and stored, critical and important tests are again repeated in the real-world environment using automated cars with other real subsystems to depict the surrounding environment, which are all controlled by the AI-core, and meanwhile the AI-core is always in the loop and learning from each and every executed test case and its results/outcomes. The main goal is to achieve efficient and high quality test and validation of systems for automated driving, which can save precious time in the development process. As a future scope of this methodology, we can step-up to make most parts of test and validation completely autonomous.