Two kinds of AI Solutions...   and we need both!

 

There are (at least!) two  kinds of of intelligent solutions. A burning business problem that can be solved using a smart / AI application. For example, why customers are leaving, why business is shrinking in certain geos or with certain product segments. In these kind of problems, we start with an issue, of course, work backwards, collecting evidences (data) and analyzing the data for the root causes and remedies. Typical diagnostics whether it is equipment or human body - fall into this category - reacting to an event or issue.
The second category is more pro-active. Even if we don't create a solution, the sky is not going to come down. However, we may be sitting on hidden magical steps and decisions where we can scale up the operations, improve productivity, prevent outages/catastrophes, make our planning more efficient, etc. Examples of such applications are price optimization, auto-analysis of documents, risk predictions, discovery of new materials - all of which need intelligence that stems from data.  Here we start from data - sometimes with an ill-defined or vague objective or sometimes explorative.  By the very nature, such applications may not be very appealing to industry, because they can still survive without it today.  And, only innovative companies will think about tomorrow and day after. So, if the funding is a crunch, acceptability of such innovative AIs will be lesser than  the first category of reactive ones where it helps to solve today's problems.
"Succeed with AI doing it backwards" is relevant when we are thinking of immediate benefits. Creative applications will take time to mature and achieve acceptance. Compare Yahoo and Google.  Who believed in Google back then. Look at the stories of autonomous driving. Did we really need it? Did it solve any immediate business problem?  No. However, now, it became part of all "modern auto software in one form or another" and exponentially influencing the value.
So, in short, we need both these approaches.
- doing backwards by finding solutions to burning problems and resolving the pain points which will be quicker successes
- identifying innovative applications that make life better  which often need blind searches for (mining)  the unknown jewels in the data, Successes here may not immediate, failures my be more prevalent, however, we need to focus on this category too to advance the technology, to achieve orders of magnitude leaps in productivity, efficiency, and quality.
Glad to discuss this further

Smarter Data Containers can help improve data protection

 It is the technology that led to the massive generation, proliferation, and sharing of data that changed even the human behavior. While it has immensely benefited many useful tools and techniques, it is obviously a double-edged sword. An example is the face recognition from images of public data to spot criminals, missing children, etc., where at the same time it is an invasion of privacy and may even lead to unforeseen criminal activity too. We have to resort to technology itself to find a solution to this dilemma between openness/sharing of data to protection/regulatory regime. Two reasons for this. One, I believe that it is difficult to predict or control human behaviors and motives. Second, any complex encryption of data, taller/thicker firewalls, and deeper hiding of data may not ensure safety as infringement by ingenuous (often miscreant) minds. Human errors and bugs in technology (like the Intel CPU architecture flaw) could be another source of windows of opportunity for misuse and unauthorized access.
While I agree that 100% foolproof data protection may a utopian dream, another approach - may be bit controversial - is to make the data more open, transparent, and trackable. Let the data be embedded with information on its credentials and the tools may look for the credentials and notify the owners before processing. That may be a better answer for so called "undesirable but not necessarily unauthorized access" such as what happened with Cambridge Analytica scenario.
In addition to trying to protect by not allowing to access the data, another way to safeguard is making the data transparent, but tracking the access such as who is downloading, who is using etc. Rather than hiding the data, if we focus on the tools and containers of data to examine whether it is authenticated to process or keep the data.
In short, let us make data more “self-aware”, develop "smart data containers", and make the tools for data processing/analyzing more “socially-sensitive” which would help us track and notify the movements of data. It is a kind of “electronic anklets” or “soft-surveillance” for data by which we are not trying to prevent the movements of data but track and monitor the movements. Not that this would prevent 100% of unauthorized access, however, it could reduce some extent of misuse because of the fear of being tracked, at the least!!!

What is the Value of Information - it depends on the context!

Everybody agrees that "Data is the most valuable asset".  But, then how do we assess the value of each piece or a collection of data or information?

Why it is important?  It is critical because organization spends money on collecting, curating, and storing data. Data can be sold. Monetizing data is a vital fundamental for many commercial / social organizations.

A piece of information may be valuable to me but absolute thrash /useless item for another person.  The information I value very much today may be completely neglected tomorrow.  It all depends on who I am, where I am, etc. The value of data comes from several contextual factors.  There are more to discuss about this.  May be you all can chime in

......

Intention emerges from Context

Imagine two scenarios as shown in the picture below. For the same question, two different answers are returned which are "magically" correct and intelligent. This requires analyzing the sentences/question/message beyond syntax, semantics but needs a good mix of "pragmatics", "common sense", possible goals and actions and the steps towards the realizing the goals. This amounts to analyzing the message/question in its situational context, including the knowledge and beliefs of the speaker and the relationship and interaction between speaker and listener.

In person to person(s) communication, one of the main factors of effectiveness is the intention behind the message. The intention is not directly inferable from the words. It is a complex combination of the context and semantics of the message. The context include the "who", "whom", "when", "sequence of events/conditions", let alone sarcasms and jokes. An intelligent person often quickly understands the intention while a novice might just grapple with the meaning of the words. The non-verbal aspects further complicate this. The dissemination of the intention of a message (question/advice/instruction, etc. ) is very critical for finding the right answer for the question, taking the right action on the instructions, etc.

Also, it is critical to understand the intention for providing better education, better explanation, and in general, more effective communication. It is often not difficult to find an answer, however, knowing what answer the user is expecting clinches the deal, improves customer satisfaction, and avoid confusion and misunderstanding. It also provides valuable information on safety and protection too from any of adversary intentions of dangerous people/criminals.

When we are enabling the computer to be more and more cognitive, it is necessary to go beyond typical NLP analysis/understanding if we want the cognitive application to behave intelligently.

A simple question like "Is it going to rain today?" can have different connotations based on the context that helps you understand the intention. It may be to switch off the sprinklers or cancel an outdoor game. It may be as simple as what kind shoes one should wear to work, to determine the attendance for a party, simply to know what the temperature could be. It may not be possible to infer the right intention without knowing the context without the previous or subsequent questions/communications. This is where we may have to piece together several factors in understanding the intention.

The dissemination of the intention of a message (question/advice / instruction, etc. ) is very critical for finding the right answer for the question, taking the right action on the instructions, etc. Also, it is critical to understand the intention of providing better education, better explanation, and in general, more effective communication.

 It is often not difficult to find an answer, however, knowing what answer the user is expecting clinches the deal, improves customer satisfaction, and avoid confusion and misunderstanding. it also provides valuable information on safety and protection too from any of adversary intentions of dangerous people/criminals. Deriving the right intention is of prime importance to understand the message and react accordingly.

Useful, marketing communication, threat detection, education, etc. May be detecting the intention from the words and actual intention can also indicate possible deficiencies in cognition too once this method is fully developed which can be helpful in psychological and medical analysis of cognitive development in probable patients.

Why understanding the intention is important?

Predicting upcoming behaviors so that better product/service/price can be offered

Anticipating upcoming questions so the answers can be found more accurately and quickly

Detecting criminal activities so that appropriate preventive actions can be taken