idea 1
Humanity produces 6,000 tweets, 40,000 Google searches, and 2 million emails. By 2020, global web traffic will surpass 2 zettabytes per year.
differentiate web traffic from real data.
Does a google search count as a document? rather information.

idea 2
Data not connected (or not saved in the cloud) or not used by Bigdata or AI : a lack or a flaw to be filled.
Intangible information : human / machine interaction (response), feelings, common sense, documents on non-scanned paper, any object that has meaning (semantic or transactional)

idea 3
Servers that collect info (Data Crawlers) for the big data, how they manage difference between human responses (or requests) from those of the bot (or other AI).
Problem of security and data integrity (example use of captcha)

idea 4
Do the large firms using and offering AI, machine learning and (chatbots, machine learning, Cognitive Computing, etc.) have they developed enough algorithms to answer all Questions ?
Does their cloud (bigdata) collect all necessary info ?!

idea 5
Most applications are some kind of human assistant or replacement :
large target of customers ---> large volume of response or offer ---> requires large volume of processing ---> we use the AI ​​(speed plus cloud and big data availability)

idea 6
In some cases, assistants were reserved for people with disabilities.
voice assistant, fingerprint, autonomous car, self parking cars, etc ...

idea 7
AI and Machine Learning rely heavily on Big Data.
The data allow the development of predictive models.
Who will define these models, the humans or the program itself.
Problem of prediction.

idea 8
Availability of bigdata platforms offers on the cloud will pose a security issue.

idea 9
All of these advances are based on connectivity. If you don't have a connection, you can't take advantage of cloud, computing power, etc.
Design an autonomous intelligence, without the need for connection

idea 10
For cognitive systems : voice recognition, speech, text, images - how to interact with unknown objects (new language not listed, codified language, cipher text, etc.)

idea 11
In the future, are we allowing machines (algorithm programs) to create new algorithms and expert systems themselves (Bigdata analysis, decision making, prediction, etc.). Who will judge the quality of these new models / algorithms ?

idea 12
Maintenance management (during their lifetime) for existing algorithms / programs.
If I11 is true, then it is necessary to create control models (or checklists) for integrity, homogeneity (according to some defined ethics).

idea 13
For Deep Learning (case of supervised learning), it's only statistical algorithms, using heuristic models to train and learn the right answers (practices). With learning, algorithms will be able to make decisions.
Applications installed in our mobiles are often cookies and data collectors on behalf bigdata. Those informations would then be analyzed by algorithms. It's in no way AI, but simply a system of collecting, selecting, sorting, comparing data.

AI relies more on examples than instructions. It relies on a group of data used by the learning machine, allowing it to provide answers for categorised data it has never seen

idea 14
Firms have created a new business : maching learning and deep learning, to take advantage of the large volumes of information included in big data.

idea 15
With convolutional neural networks, Deep Learning is at the heart of computer vision and robotic vision. What about a more evolving AI that processes objects (entities, elements, concepts, ideas, thoughts) ? we go to another conceptual level which is the most suitable for communicating with humans.

idea 16
Intelligent robots or machines are just a heap of junk powered by electricity with a code that runs and is programmed by humans. It can never be INTELLIGENT. he can never think of himself.

Issues : Try to imitate as much as possible the functioning of human thought. Coding an AI capable of thinking for itself while interacting with the outside world. Create M7 context, without the need of bigdata

idea 17
How does the human mind perceive, analyze and process the data / info retrieved by these sensors / organs / senses : listening, seeing, touching, smelling, speaking ? if we succeed in imitating the perception of the senses, we will approach the mechanism of the brain. How does the brain know how to differentiate between these different sensory objects ?

idea 18
Offer autonomous and secure AI solutions (data analysis, intelligent networks, decision making, etc.) for SMEs or groups who want to keep their data secret.
It also makes it possible to keep control over this data, and not to pollute them with unreliable external data.

idea 19
Microprocessors and other information processing equipment cannot be intelligent.
How to use them ?
1 / design a thought model close to human behavior (intelligent)
2 / how to store information (BDD and DD)
3 / relations / connection between information
4 / redefine the design and operation of microprocessors

idea 20
With I19 the human brains will be the processing units. we will have as many processor or cores as brains. Bigdata data will be the information stored in the brains.