Certainly a big YES. Artificial Intelligence has potential applications in medical procedures involving robots, self-driving vehicles, farming and agriculture, defense systems, psychoanalysis of weirdos and tracking down crime suspects quickly, personal assistants for individuals with special needs and old age care, accurate and timely predictions of devastating weather patterns, laboratory assistants who conduct routine experimental testing and analying big chucks of data as you mentioned. One can also include the development of new scientific theories based on the vast reservoir of already established theories detailed in journal articles (arXiv, researchgate, etc) by a system with intelligent brain architecture. Its possible that a smart system can put together two seemingly unrelated theoretical concepts and come up with a unique and much superior idea in quick enough time, if quantum effects are harnessed to built these machine. Developing smart learning machines that incorporate highly accurate predictive power is the way to the future, and hopefully these machines can result in a peaceful, non exploitative environment that is used for the well being of all.

Artificial Intelligence tools are more accurate than traditional tools.

Several algorithms deal with specific goals as follows:

Critical success factors identification:

1. Neural Networks

2. Fuzzy Cognitive Maps

3. Genetic Algorithms

4. Bayesian Model

Project success prediction:

1. Bayesian Model

2. Evolutionary Fuzzy Neural Inference Model - EFNIM

3. Neural Networks

4. Support Vector Machine

5. Fast Messy Genetic Algorithm

6. K-Means Clustering

7. Bootstrap aggregating neural networks

8. Adaptive boosting neural networks

Renat,

While there is an answer to your question, the first statement "scientific data that can probably solve all the problems of mankind" is a huge assumption that I would be careful in making either for or against. The second point to clear is that big data does not always equate to good data or relevant data. Also take into account Bonferroni’s Principle [1].

Having said the above, the answer is not likely unless more advances are made in symbolic-subsymbolic integration. The caveat is how to extract the principles from raw data when they are implicit in most cases. Also, the design of these principles would have to be predetermined and then examined to see if they can be empirically mapped to the data.

If you design an AI that could in principle derive those principles from data it would more than likely be biased, based on the designers preconceptions of what he thinks are those principles. Techniques such as clustering and other ML techniques are not effective in this space. Latent semantic indexing might be a good place to start to get some commonalities, but from there to deriving abstractions to get principles is a long way off.

[1] Mining of Massive Datasets 2nd Edition by Leskovec , Rajaraman, and Ullman pp. 5-7

This is a very difficult question to answer. First at all we must define "the research" and second what do we mean by "Artificial Intelligence". AI is not a science in the traditional sense. It is a collection of approaches modeled inside a field of one science (mathematics, physics, biology .....social sciences......) or inside a fiels of two or several sciences (examples : genetic algorithms, convolutional neural network). These approaches raise problems and try to solve them. Once a model proposed, it is transformed in an algorithm and this algorithm is injected on the machine. If the machine can take in charge a number of tasks more and more "sophisticated" that human brain can solve, it become an artificial intelligence system.

So, it's not only the problem of data, it is the problem of the model, the approach, the algorithm and its running on the machine.

I think that is the epistemological path for an AI system to be an AI system.

As for the research, it is a research on all vertex of above path. The difference is that there are different types pf research. The research in mathematics, are different from the research in computer science and eve in the different sub-field of computer science.

This is my position.

(1) Interesting views on AI from researchers above, but judging alone from the interests of students considering a research career, AI may well be unstoppable. It is not clear if machine can perform tasks "more sophisticated than human brain can solve" as mentioned in an earlier comment. Can Artificial intelligence supersede human intelligence? This raises the question of AI safety and ethics etc. At this stage, the healthy human brain is irreplaceable, one of the most superior system in existence.

(2) More likely augmented intelligence which involves the cooperative and complementary participation of predictive machines under minimal human supervision and control will allow strategic tasks to be performed with efficiency and high speed. Many boring tasks will be eliminated.

(3) Lastly I wonder if you constructed a AI system using the remarkable apparatus consisting of the sulphur bacteria or photosynthetic complex with their adaptation marvels, would it still be called "artificial" intelligence?

Certainly if we consider AI for handling complex problems in contrast to traditional mechanisms it is a big YES. There are a few parameters that can be overlooked if observed manually, but beauty of AI is that it would consider every minutest details as trained. Moreover if the agent program is in self learning mode it is far better than human brain.

Renat, we have indeed large data today, but I doubt that they can solve all problems of mankind. Most likely, we do not have data to solve all these problems. Even worse, we may not know what to measure and how, and how then to process the data. These are tasks of scientists.

I know that scientists can work with Big Data. There are even big data conferences. Suppose that you have some characteristics for the whole population of 7 billions, like age, sex, income, profession, etc. Any person cannot even read it during the life. But it is possible to assign statistical algorithms to work out what we need. Computer will work, but according to an algorithm assigned by a scientist.

If the complexity of interaction between data becomes larger, so that scientist cannot work out a good program to deal with that, I doubt that AI can do it better (today and even in visible future). AI can try to use several algorithms and to find optimal, applicable to those big data. It can also be asked to learn and to adapt algorithm. But it cannot set a correct objective when a person is unable to do it.

If you know some examples when AI was able to get some result from the scratch where scientists did not know how to approach it, please let me know.

Yes, Because of its outstanding features..

The future of human life is certainly determined by the innovative trends now available in AI. Definitely, these trends offer fantastic and unprecedented advantages to the scientific research and development. The remarkable achievements in AI have enormously influenced and altered the general educated opinions so that the by products resulting from it in the form of various statistical software have revolutionized all aspects of doing scientific research.

Today Artificial Intelligence plays vital role in Computer field. Artificial intelligence (AI, also machine intelligence, MI) is intelligence demonstrated by machines, in contrast to the natural intelligence (NI) displayed by humans and other animals. In computer science AI research is defined as the study of "intelligent agents": any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.Colloquially, the term "artificial intelligence" is applied when a machine mimics "cognitive" functions that humans associate with other human minds, such as "learning" and "problem solving".

The scope of AI is disputed: as machines become increasingly capable, tasks considered as requiring "intelligence" are often removed from the definition, a phenomenon known as the AI effect, leading to the quip, "AI is whatever hasn't been done yet." For instance, optical character recognition is frequently excluded from "artificial intelligence", having become a routine technology. Capabilities generally classified as AI as of 2017 include successfully understanding human speech, competing at the highest level in strategic game systems (such as chess and Go), autonomous cars, intelligent routing in content delivery network and military simulations.

Artificial intelligence was founded as an academic discipline in 1956, and in the years since has experienced several waves of optimism, followed by disappointment and the loss of funding (known as an "AI winter"),followed by new approaches, success and renewed funding. For most of its history, AI research has been divided into subfields that often fail to communicate with each other.These sub-fields are based on technical considerations, such as particular goals (e.g. "robotics" or "machine learning"),the use of particular tools ("logic" or artificial neural networks), or deep philosophical differences. Subfields have also been based on social factors (particular institutions or the work of particular researchers).