Defining AI's Future - Less Or More Data
With the rapid plunging of many firms into AI's technology, a belief that AI will be using more data has risen. It is therefore important to define the future of AI, will it be more data or less data? Will it be relying more on data? The answer is no.
AI's FUTURE, LESS DATA? YES!
First, companies looking to integrate AI with their technologies should know that in the nearest future, AI will be relying on less data than they do now. That is, with each coming year, AI uses fewer data. In fact, it might come to a point where AI will no longer use data. These machines will become more than just an artificial creation, become more intelligent and rely more on their intuition, just as humans! Isn't that an exciting prospect?
The reliance on bottom-up big data will be considerably diminished, while the top-down reasoning that closely resembles the manner humans think and do tasks will be more employed. This intuitive ability will give a surge to the use of AI, and also its handling of facets for which it was previously unsuitable for.
WHY THE LESSER RELIANCE ON DATA?
Since inception, AI has progressed through deep learning and machine learning using bottom-up data to train its capabilities. However, this system has a number of hitches because the machines are rendered clueless in situations which they do not have sufficient data on. An autonomous that has been trained on data for many traffic situations can perform adequately in many casual situations but will have no clue in certain edgy situations. This situation may be children in Halloween costumes weaving through the streets at night or any uncommon circumstance.
There are certain system weaknesses too. The iPhone X does not recognize morning faces or puffy looks when you first rise in the morning. Neural networks have also won chess competitions but a slight alteration to the image disrupts the AI's system and cause a wrong identification. Sometimes, the system may use its 'high confidence' feature to give you a wrong identification.
Businesses using AI also have some constraints and may be susceptible to manipulated outputs since there is no clear explanation for how the AI gives outputs. Furthermore, many companies do not have the voluminous feed of neural data required to operate AI effectively. The use of citizen's data is limited in order to protect privacy.
In fact, the European Union has laid down the General Data Protection Regulation (GDPR), a protocol that requires governments to fulfill certain difficult conditions before citizen's data can be used. Again, these systems are unclear, hence there is no sure explanation for how they use inputs to perform tasks. A clear example of this can be seen in the 2016 USA presidential election.
Soon, in the nearest future, AI will be running on the top-down system. With human intuition, there will be more efficiency, flexibility, and intelligence. These systems will also handle sticky unexpected scenarios faster than a human. Some companies have already begun the use of these more natural systems. In the soonest future, AI may be our closest relation, moving above the ape in the pecking order.