Computers may be powerful, more accurate and with a huge power to compute and handle complicated tasks, but it takes Human Intelligence to build them and make sure they are function as intended, HI can therefore only be augmented but not replaced.
Ehm Kannde
In 2012 as I was turning out from senior high school and I was a little bit confused as to where to go next. In the end I found myself in a medical college in a discipline that was just starting out in that school and relatively new in the world. One of the big questions on the minds of many students was that of job security since the profession was new, they were wondering if there would be readily available jobs when we graduated.
As time progressed, there was the growing fear and panic among students with the growing automation in Laboratory Medicine industry. Now machines can count blood cells, differentiate cells and identify disease patterns more accurately than humans could possibly do. Many had the view that the more jobs machines are able to do the more we will sit at home jobless.
They were not wrong to think this way because it has happened several times in history where workers had to be laid off when robots were made do the same tasks.
To add more insults to injury, analyst PWC in a report stated that 40% of jobs will be gone by the year 2030 due to artificial intelligence.
This presented a political, economic and social dilemma in a sense that AI has helped solved so many problems and still helping so how do we balance this and the growing job insecurity?
Many enthusiast of the AI age referred to as the forth revolution often stated that machines were only meant to augment HI and not to replace it but this does not come to play when more workers get laid off!
In trying to find a solution to this, Bill Gates proposed that taxes should be imposed by governments on operations undertaken by automation. It is obvious that companies would prefer Artificial Intelligence (AI) to Human Intelligence (HI) because of several reasons but chiefly to maximize profits and taxing them will limit automation allowing people to keep their jobs.
However, this is yet to be actualized as more and more jobs are being taken as AI advances.
In the next 15 years, machines are expected to be able to drive cars, replace soldiers, work on the factory line building components, provide comprehensive customer service, run our financial services, translate and interpret text quicker than humans and and (the list goes on and on)."
The metro, UK
More problems?
Despite the general fear of machines replacing humans at work, the relevance of AI cannot be overemphasized and it's shown by tonnes of research that is being done on advancing the knowledge on robotics by the year. The AI economy is continuously growing and estimated to reach a staggering 15.7 Trillion market capitalization. This sounds great but there are still tonnes of problems facing the development of this fast evolving industry. These problems can be put into 3 main categorizations;
Data processing pressure: It is said that the human brain is the most complex structure in the universe mainly because of our intuition. This means that AI involves the use data sets and annotations that will allow it to function and this data sets is just too expensive which limits individual and companies with limited budgets.
Diverging of tasks: Ever wondered how your smart phone is able to adjust automatically to the amount of day light? Well, this requires learning from the environment and how it interacts which calls for integration of other systems making AI much more complex making AI project hard to complete.
Cost of computational power: AI requires huge computational power as many algorithms and processes involving terabytes of data has to take place simultaneously yet in complete coordination. It is difficult to provide this computational power without huge budgets.
These 3 problems has led to the creation of an environment where only powerful firms and governments is able to utilize these technology enshrouded in a cloak of secrecy which only leads to the rich getting richer and the poor even poorer.
The Effect.Ai as the solution
What is Effect.Ai?
This a decentralized network built on blockchain technology for the AI industry.
The Effect Network is out to provide the solution to the problems stated above by way of leveraging the power of community which is made possible by blockchain technology.
In order to meet the growing demand for AI, Effect Network aims to provide the environment that will allow individual developers and small business to get unrestricted access to data sets, task diverging algorithms and vast computational power without any fees from intermediaries.
How would Effect.Ai provide this solution?
One step at a time, this is a living and breathing ecosystem that is divided into three phases;
Phase one: Effect Mechanical Turk
Effect.Ai will bring Mechanical Turk to blockchain during this phase of its development.
Mechanical Turk is meant for developers from any part of the globe to access 'workers' for 'tasks' that demands human intelligence (HI). Developers can access a vast network of people who will help in 'teaching' their AI projects. At this stage, any person can earn tokens by completing tasks for 'requesters' making it a marketplace of "machine education". Individuals get to contribute to the development of this project while earning in the process.
Someone may ask why a company would be motivated to chose Effect Mechanical Turk over Amazon. Well, request from developers is free as it is a peer to peer network. This will enable project owners pay 'contributors' better and that will not only ensure quality job but will also aid in 'workers' meeting timelines in a generally more efficient manner.
Tasks may include Image Classification, Transcribe Audio and Data Labeling
Phase two: Effect smart market
This phase involves the release of a decentralized marketplace where AI related services is traded. This may be services built using phase one of Effect.Ai Mechanical Turk or any AI application or algorithm developed outside of the network. This services are listed where any developer can easily purchase, paying directly to the owner in order to expedite the development of their own applications.
There comes a time when bots can complete some 'tasks' that were initial done by 'workers'. Because of the efficiency bots may bring, the network has decided to make it possible for the bot owners to earn on completed tasks .
Phase three: Effect computational power
With phase 1 and 2, the algorithms are still running on centralized servers making it impossible to achieve a fully decentralized state. Phase 3 brings the final solution by distributing data across a network of computing power which will run on different nodes without a single point of failure. This phase is termed as the 'Effect power' phase.
This completely decentralized ecosystem has so many long term sustainability related questions. The Effect.Ai team have worked so hard to provide solutions to all anticipated future issues in order to fulfill sustainability agenda. This has called for components of governance meant to maintain a strict economic structure, reputation score meant to encourage provision of only the highest quality results on tasks and the galaxy pool which is meant to prevent undue manipulations by short term minded investors and to make sure there is a demand for the EFX tokens. Details on this system can be found on the Effect.Ai Whitepaper
Diagram of the Governance Model and construction of the Effect.AI
Galaxy Pool
Use case:
- Ama has been working a project for sometime now, she intends to build an app that will recognizes face. She wants to implement this into home security systems so that access will be allowed based on facial recognition.
Ama have developed the interface and she now wants to teach her algorithm by showing as many faces as possible so it can know how to individualized them.
Ama can accomplish this by listing the request with the Phase one of Effect.Ai. Workers inputs images of different faces in a data set and Ama compensate them with EFX tokens.
Now Ama's algorithm now have enough images on the data set to function properly. The second phase (Effect Smart market) allows her to list her application for sale on the platform. In case Company A is producing security systems and finds Ama's algorithm powerful enough, they use it and compensate Ama which means Ama gets to earn some EFX anytime here algorithm is used by any developer.
Ama's algorithm is currently hosted on a centralized server and with growing demand this brings more computational requirement. With the their phase of this innovation, her algorithm is distributed across a global network of vast computational power guided by smart contracts made possible by the Neo blockchain technology.
You can read more about the token sales HERE
You can also watch the video below for more information:
If you still have any further questions, please check out the Effect.AiFAQs
You can also contact the team in their extremely active Telegram Channel, on their Facebook page or by following them on twitter
This is my submission to the Effect.Ai sponsored contest with @originalworks.
effect.ai2018