IA the solution the AI control problem. Research Proposal

CAN IA BE THE SOLUTION TO THE AI CONTROL PROBLEM?
Name:
Institution:
Course Code:
Date:
List of Abbreviations
IA – Intelligent Augmentation
AI – Artificial Intelligence
Abstract
It is possible that AI systems can go berserk and pursue different goals than those intended by the creators. Though AI is in its infancy, experts have warned that if it is not developed safely without a ‘kill switch’, it could be apocalyptic. There have been several studies to avert such catastrophic outcome. This research project aims to use secondary research to analyze findings from other sources to ascertain whether Intelligent Augmentation (IA) is a viable solution to control Artificial Intelligence (AI). The study will explore the developments in transhumanism, cybernetics, bionic limbs, implants to ascertain whether they are good enough to support IA integration with humans.
Introduction/ Background
Humans have been attempting to alter their physicality and any limitations they have such as strength, life expectancy, health conditions like Parkinson’s disease etc. On the other hand, AI systems have inherent problems stemming from the data used for training and how the system is designed. Any anomaly and or an error in creating the AI or in the data used for training will translate to problems with the AI. However, advances in medical and human computer interaction have also provided new ways to use technology to solve problems related to human body. Micro chipping, RFID (Radio Frequency Identification) implants, bionic limbs, cybernetics, transhumanism etc have redefined how human body and computers can interact. Today, some people have bionic limbs which functionally nearly the same or better than their biological ones CITATION Blo19 \l 1033 (Bloomberg Quicktake, 2019), others have implants that treat previously untreatable conditions like Tourette’s syndrome CITATION Eli17 \l 1033 (Strickland, 2017), paraplegic people have taken their first steps with help of spinal cord implants after spinal cord injurie CITATION JoK18 \l 1033 (Khan, 2018)s. Therefore, rather than designing AI systems that are entirely mechanical, can human body be interfaced with machines to create a stronger hybrid? The main question is, is it possible to improve human intelligence, productivity and performance by use of technology? It is clear that the line between what is considered machine and what can be considered purely human has become blurred. Advancement in technology has shown potential to address these problems through transhumanism, bio-hacking, IA (Intelligent Augmentation) and AI (Artificial Intelligence).
Statement of the Problem
As the line between sci-fi and reality becomes blurrier, questions of whether AI is controllable arise. At the moment, nearly all AI systems have an inherent problem, they are as good as the data which trained them. In most cases, the data is biased and hence the AI system is also biased. AI systems barely still have the intelligence of a 6-year-old though it is fast improving CITATION Zoe17 \l 1033 (Chong, 2017). However, even if the capabilities of AI improve, there is no way to guarantee future AI systems will safeguard the best interest of humanity. Technology is still in its infancy stages and as we head towards general artificial intelligence, questions linger about how we could control our creation. Pessimists fear that it could wipe out humanity or enslave us. Elon Musk has warned AI developers to be careful not to create a machine that can become immortal dictator CITATION Rya18 \l 1033 (Browne, 2018). Most AI systems in the world today have inherent problems that show the importance of controlling AI. For example, even the best autonomous car systems have shown bias or have still caused accidents. Autonomous cars are some of the best application of AI but despite the large data sets for training these systems, they are still fooled by small things on the road. Rather than replacing the driver altogether, what if the driver’s senses are heightened by intelligent augmentation? This would largely reduce the possibility of accident as well as ensuring that humans are still in control of the world. Intelligent augmentation shows that AI could be managed and controlled in a way that we could reap its benefits without risking the possibility of having it rule over humanity. Rather than entirely designing systems that are entirely machines, we integrate the AI with humans to improve their performance and productivity.
Purpose of the study
The aim of this study is to sift through existing literature to find out if IA can be the answer to AI control problem.
Research Questions
1 To find out whether IA t...