IT is becoming increasingly apparent to all, especially among scientists, investors, and entrepreneurs that machines will soon become smarter than humans.
In a recent event held in Berlin, Germany, artificial intelligence (AI) experts discussed the “Rise of AI” in front of an audience of 160 people.
Among the things addressed was a quote from science fiction author William Gibson, stating: “The future is already here, though it is unevenly distributed.”
. Factors facilitating the growth of artificial intelligence
First, there is a lot of progressive machine learning, owing largely to advancements at Google, Amazon, Apple, and Facebook.
The idea of machine learning itself has been around for a while, but today’s advantage is the much stronger computer processing power which has made it possible to achieve a much smarter AI than in the 1960s or 70s.
Artificial intelligence is largely dependent on the processing power of computers. With modern technology, it is much easier to achieve results that were not possible before.
Machines have a better understanding of how humans write, speak, or use pictures, which has significantly enhanced the process of human-machine communication.
Third, there is data mining: corporations and many internet companies have been collecting data for the last 10 years or so.
These data-driven companies now need machine learning or AI tools to figure out this data and implement it. Without this data, there would be no learning fodder for machines or AI.
Fourth, the internet was not available decades ago. The internet provided the means for AI.
. Applications of AI
With an AI enabling environment, there are more startups working in the AI space, resulting in more trained robots to handle a myriad of duties that were previously performed by humans.
AI is good at specific tasks today, such as:
- Playing games – chess, cards, etc.
- Flying drones
- Driving cars, planes, boats – Tesla, Google, etc.
- Organizing meetings
- Booking plane tickets
- Shopping online
Chatbots are essentially taking over roles that were previously assigned to personal assistants. AIs can also be used to:
- Write songs
- Draw pictures
- Trade stocks – robot advisors
- Write emails
- Facilitate Instant medical diagnosis
Many industries are also benefiting from the ability of AI to optimize complex processes and solve problems that were impossible to solve due to large amounts of data and variables, collected over years of experience.
AI tools can go through this data and derive solutions that humans cannot see, like in:
- Production cycles
- Supply chains
- Manufacturing processes
. Are AIs better at everything?
AIs are better than human beings at very specific tasks that they have been TRAINED for.
Previously, it was pretty expensive and time-consuming to train machines to be better at some things than humans. But today, with the internet, data, and progress in machine learning tools, it has become cheaper, easier, and faster to train machines.
So, developing an application that is better than a qualified worker at a specific task can be done in months, weeks, or even days. Moreover, machines are capable of working 24 hours a day without fail.
This, unfortunately, increases the risk of mass unemployment, as more jobs will be lost to technology, compared to those created by it.
Companies are looking to automate different processes, from driving to medical diagnosis to legal counsel; so it is possible that truck drivers, doctors, and lawyers might get replaced by intelligent programs in the near future.
With Uber introducing self-propelled taxis in Pittsburgh, US, some taxi drivers may have to find alternative ways to earn a living.
After all, the robot on wheels will make the driver’s role redundant while making taxi rides cheaper and the preferable option.
. Should humans compete with machines?
In some areas, machines are already better at certain tasks than humans, which is not necessarily a bad thing. But in what ways are humans unique?
According to experts, people argue that machines will never be like human beings, especially in regard to the human thought process.
And this single factor will keep them from taking over everything. But this is not necessarily true. Robots can be trained to do just about any task and be better at it than a human.
They can be integrated into every area, including manufacturing processes and stock trading, where they are fed with data from the many work hours, and left to figure out how to optimize the processes.
Huge amounts of data and variables that humans may find hard to analyze are not a problem for machines, making them better at any trained process.
Machines can be trained to be creative and even to dream; they just don’t do it the same way as human beings. Machines can even identify trends; however, they lack conscience compassion or emotions – love, greed, hate, ethics, humor, soul, curiosity, insecurity, DNA, etc.
Humans need machines to increase productivity and output. Since industrialization, people have enjoyed more free time and have become happier and healthier. Life expectancy has also gone up, and the same can be expected with AI.
. Can AIs take away human resources?
For AIs to take resources from human beings, they would have to “wake up”. This means that they would have to acquire consciousness and intelligence beyond that of human beings.
But AIs becoming smarter than human beings is not too farfetched. Here’s why:
1. The substrate that AI runs on
Artificial Intelligence “runs on Silicon”, which is a $300 billion a year industry. To bring this figure into perspective, the Hollywood industry is at $50 billion a year, while the video gaming industry is about $100 billion a year.
The semiconductor industry is growing drastically and following Moore’s law. There are PCs, smartphones, networks, electronics in cars, etc.
2. Billions going to AI directly
Google, Facebook, and other top tech/internet companies are buying startups for millions of dollars to leverage artificial intelligence and increase click-through revenue.
For instance, in early 2014, Google bought UK AI startup DeepMind for $650 million. The firm specializes in machine learning, systems neuroscience, and advanced algorithms, and is claimed to reduce Google Data Centre cooling bill by a remarkable 40 percent.
More recently, Intel spent $400 million to acquire a deep learning startup called Nervana Systems.
3. DAO Infrastructure coming of age
The concept of DAO – Decentralized Autonomous Organization is an idealistic result of a crypto-tech revolution.
DAO is on a path to an evolutionary sequence with each stage building on the functions of the previous one, and will eventually get a point where autonomous agents can do their work through artificial intelligence or smart programs.
These stages are:
- Participative – where users voluntarily and independently take part in loose tasks
- Collaborative – where users collaborate and add value towards reaching a common objective
- Cooperative – where users expect some shared gains in return
- Distributed – propagation of these functions begins by multiplying them across a wider net
- Decentralized – more scalability is reached by increasing power in the edges
- Autonomous – smart programs, autonomous agents, and increased levels of AI and AI algorithms will provide self-sustainability in operations and value creation at all points of an organization.
Still, a new DAO is like a startup with many assumptions being made before it can become sustainable, but it will be a major contributor to AI.
4. No fundamental roadblocks
AI is just as creative as human beings, plus it has been proven that robots can be trained to do just about anything a human can do, and even better.
If AI promotes productivity, speed, and profits, it will be more readily adopted.
. The Future of AI
In the not so far future, not necessarily in a 25 years time span, intelligent machines will be super-smart and probably capable of ruling over people.
The technology will have reached a level where it can optimize itself further, giving rise to the theory of technological singularity: the extreme acceleration of progress through machines that continuously improve themselves.
According to this theory robots and applications will eventually be smarter than their inventors in any way imaginable.
Some people may dismiss such claims as fantasies of science fiction, but more and more companies are investing in innovations aimed precisely at AI and robots.
In the last five years, investment in AI startups has increased six-fold. In 2016 alone, venture capital firms have invested in over 200 startup companies dealing with artificial intelligence.
Most of these companies are developing programs for specific, highly localized programs, like self-learning programs that diagnose diseases, or software to detect credit card fraud.
Some are even working on the concept of mind uploading, where users transfer their own brain to a computer storage media. This technology may be available by 2030.
. But would humans be willing to upgrade as well?
During the Berlin conference, the attendants were asked to imagine technological progress over the next 25 years and then surveyed to find out whether they would be willing to upgrade to increase their intelligence, health, and happiness.
Only a handful claimed that they would not change a thing, while most admitted that they would be willing to use neuro-enhancement, biomechatronic body parts, neural implants, life extension with cryonics, or upload their mind to a computer.
People are already relying on technology like smartphones, which are more or less an extension of the brain. People also depend on software to recommend products, control vehicles, and write and send messages. So there is nothing to deter them from mind uploading if it adds value to their lives.
Assuming that humans observe the world at one second per second, then moving your conscious activity to the medium that performs at a more efficient level than the brain would mean that you would be able to observe a lot more per second. The world would literally slow down, allowing you to do more and experience a lot more in the same amount of time.
But in a real sense, humans perceive a lot of things in a single second, yet tests have found that human response time has a minimum of 50 ms. So, if you were able to perceive a lot more per second, you would also need to respond a lot faster for that extra information to be meaningful.
. Enhancing your mind
Science has made it possible to simulate simple neuron connections to thousands or millions of other neurons through synapses.
In a study published in the Journal Frontiers in Human Neuroscience, subjects who received brain stimulation through electrode-embedded head caps exhibited better piloting skills and performed tasks 33 percent better than the control group.
The researchers studied electric signals in the brain of a trained pilot and fed that data into inexperienced subjects as they learned to fly an airplane in a realistic flight simulator.
Brain simulation is based on the idea that when you learn something, your brain changes physically, and the study provided some proof to this.
Mind enhancing is not just a way for humans to catch-up to increasingly smarter robots, but also an opportunity accelerates economic growth.
If human minds can run on substrates that are then accelerated by computing power, humans would be able to spin off mature progeny with superior intelligence to facilitate economic growth.
But for mind uploading to be a possibility, three conditions must be met:
First, it must be possible to incrementally replace a brain with functionally identical implants while retaining its identity and fundamental characteristics;
Second, the computational capacity of the human brain must be a rational number, no more than 10 – raised to the power of 19 ops/sec; and
Third, computers that fast must have been created.