Chapter 1 AI at Present
This is our first chapter for the AI Professional course.
Artificial Intelligence, commonly called AI, has been on everyone’s lips the past years. Nowadays, AI became such a “hot topic” that it is impossible to avoid media coverage and public discussions about AI.
While people often picture AI in their mind as robots with some human-like characteristics, AI can encompass anything from Google’s search algorithms to automatic question answering systems like IBM Waston, to autonomous weapons, or self-driving cars.
In this chapter we will be looking at AI from a top-level perspective, by asking the questions: what does AI mean? What are the common applications of AI and how is it applied in the business world and the public? and what impact is it having? These are tough questions, but important ones to answer.
1.1 What is AI
Artificial Intelligence, or “AI” could mean different things to different people. Nowadays if you went looking for a definition on AI, or introduction videos or general articles on what is AI, you would easily find thousands. Indeed, defining Artificial Intelligence is not an easy task as there is no official and commonly agreed definition we can refer to, even among AI researchers. Not only that, AI is an ever-evolving term that its meaning keeps changing as the field and related technologies continuously evolving.
We hereby, look at a couple of definitions on AI:
- Wikipedia’s entry on AI
- Computer science defines AI research as the study and design of “intelligent agents.” More elaborately, AI is characterized as “a system’s ability to correctly interpret external data, to learn from such data, and to use those learnings to achieve specific goals and tasks through flexible adaptation.”
- English Oxford Living Dictionary
- “The theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.”
- Encyclopedia Britannica
- “Artificial intelligence (AI), the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings.”
A few other important definitions of AI could be found here: The Key Definition of AI
| # AI is essentially problem-solving software that mimics the human brain. It examines a problem, seeks to determine what is occurring and then learns from what has happened. “It’s recognizing your face in a photo; it’s translating a menu on your phone; it’s using computer vision in self-driving cars to see what the other cars and obstacles are,” says Bloomberg tech reporter Ashlee Vance. Source: https://www.morningstar.in/posts/49472/ai-changing-lives.aspx |
DISCUSSION
- What are the common aspects and differences in the above definitions of AI?
- Did you find anything missing from the above definitions about AI?
- Could you find any other meaningful definitions of AI?
- What does AI mean to you?
1.2 History of AI
The term Artificial Intelligence (AI) was first coined by John McCarthy in 1956 when he invited a group of researchers from a variety of disciplines including language simulation, neuron nets, complexity theory, etc. to a workshop held at Dartmouth College. The purpose of the workshop was to discuss what would ultimately become the field of AI. The proposal for the conference included this assertion: “The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can be so precisely described that a machine can be made to simulate it.”
Read about the history and timelines of Artificial Intelligence:
- Brief Timeline of AI – Live Science
- The History of Artificial Intelligence – BLOG, SPECIAL EDITION ON ARTIFICIAL INTELLIGENCE
- History of artificial intelligence – Wiki [Optional]
1.3 Opportunity and Benefits
No double, in recent years AI has gone from a science-fiction dream to a critical part of our everyday lives. Intelligent machines nowadays are able to sift through and interpret massive amounts of data from various sources to carry out a wide range of tasks. For instance, AI has enabled us to analyze high-resolution images from satellites, drones, or medical scans such that we can improve responses to humanitarian emergencies, increase agricultural productivity, and help doctors identify skin cancer or other illnesses, etc.. They’ve already changed many aspects of our lives: these days we use AI systems to interact with our phones and speakers through voice assistants like Siri, Alexa, and Google; we use our face to unlock our mobile devices without any touching; we could drive cars that interpret and analyze their surroundings themselves; in online shopping, websites like Amazon monitors our browsing habits and then serves up products it thinks we’d like to buy; and even Google decides what results to give us based on our profiles and search activity. In truth, AI is touching our lives far more than many of us could have realized.
We hereby look at various applications of AI in real world.
1.3.1 Virtual Assistant
Virtual assistant (aka. Personal assistant) is probably one of the AI applications that are closest to our daily lives. “An intelligent virtual assistant (IVA) or intelligent personal assistant (IPA) is a software agent that can perform tasks or services for an individual based on commands or questions.” – “Virtual assistant” from wiki. Virtual assistants typically perform simple but necessary jobs for end-users, such as adding tasks to a calendar; make and receive phone calls; providing information that would normally be searched in a web browser (e.g., get directions, hear news and weather reports, find hotels or restaurants, check flight reservations); or controlling and checking the status of smart home devices, including lights, cameras, and thermostats. It has undoubtedly changed the way how we live, how we work, and how we communicate. Have a look at some existing AI-powered personal assistant below.
Google Duplex
1.3.2 Public health (automatic diagnose systems)
There’s a lot of excitement right now about how AI is going to change the complex world of healthcare. AI tools have been used to support human practitioners to provide faster service and more accurate diagnoses. With the help of recently developed advance machine learning techniques, we are also able to analyze data to identify trends or genetic information that may predispose someone to a particular disease. No double, AI and machine learning are transformative for healthcare workers and patients. Watch the following video on how AI is changing the overall healthcare domain as well as the role of doctors.
You can also read the following articles for a list of AI applications in the healthcare domain.
- 10 Promising AI Applications in Health Care – Harvard Business Review
- Artificial intelligence in healthcare – wiki
1.3.3 Autonomous control systems
Autonomous vehicles
The Crash Reporting Sampling System (CRSS) reported that more than 90% of car crashes involve human errors, however, with the help of AI, we are on the verge of dramatically improving that statistic. Artificial Intelligence (AI) gives cars the ability to see, think, learn and navigate a nearly infinite range of driving scenarios, making them self-driving cars.
PC Mag defines a self-driving car as “A computer-controlled car that drives itself.”
UCSUSA states that self-driving cars are “cars or trucks in which human drivers are never required to take control to safely operate the vehicle. Also known as autonomous or ‘driverless’ cars, they combine sensors and software to control, navigate, and drive the vehicle.” Watch the following video for a brief introduction to self-driving cars, as well as the evolution of vehicle automation.
Following is an interesting dynamic visualization outlooking the future of road conditions with autonomous vehicles.
Industrial Robotics
At present, industries have been automating many different sorts of repetitive human tasks, content management, process workflows, data capture and business decisions with flexible, integrated AI capabilities. Industrial robots are a big part of it. According to the International Federation of Robotics (IFR), in the year 2015, an estimated 1.64 million industrial robots were in operation worldwide. For example, take a look at the following video about Amazon warehouse robots.
1.3.4 Cyber Security
Crime prevention
These days, companies and cities all over the world are experimenting with AI and machine learning tools to reduce and prevent crime, and/or to more quickly respond to crimes in progress. For example, police forces in the UK are already using AI and machine learning techniques to predict where and when crime will happen, and deploy officers to crime hotspots to prevent it. The idea behind many of these projects is that crimes are relatively predictable, through mining a massive volume of data to find patterns that are useful to law enforcement. Uses might include spotting burglary patterns, identifying potential suspects or preventing the collapse of rape trials by analyzing emails or text messages. Overall, software can help to solve crimes by quickly analyzing information from databases and surveillance systems. And no doubt, computers could do this sort of examination far more quickly than humans.
For instance, PredPol is one of the companies use big data and machine learning to try to predict when and where crime will take place. Currently, their system is being in several American cities including Los Angeles, which was an early adopter. Watch the following video from Dr. Jeffery Brantingham (co-founder of PrePol) explaining the science behind predictive policing.
Biometric ID
Last but not least, AI has also successfully led its way to the application of biometric identification and recognition. The following article from emeRJ provides a great review of the recent AI applications in Biometric solutions for Security.
DISCUSSION
- What AI applications are you using in your everyday life?
- Above is far from an exhaustive list of AI applications. There are a lot more real-world applications of AI that are not listed here. Can you think of any?
1.4 Costs and Risks
Artificial intelligence (AI) is changing societies and economies around the world. However, despite all the benefits that AI bring us, the transformative power of AI, however, also comes with challenges, ranging from issues of privacy, bias, transparency, trust and security, to concerns about displacing jobs and exacerbating inequalities, etc. The more we study and develop artificial intelligence, the clearer it becomes that this massively powerful tool comes with a great deal of responsibility.
Privacy Concern
For instance, we’ve talk about virtual assistant in Section 1.3.1. There are serious consumer privacy concerns about such products (e..g, Amazon Alexa and Google Home), because these tools indeed require large amounts of personal data. They retain voice interactions and personal information to improve the user experience, and they are always “listening” in order to respond to voice commands.
Source from The Intercept https://theintercept.imgix.net/wp-uploads/sites/1/2018/11/Intercept_Echo_v2-3.5MB-2-1542062294.gif
Fairness
Like humans, technologies can make mistakes, and often, they display unfair bias against people of color, gender, race, etc.. Take another example, Correctional Offender Management Profiling for Alternative Sanctions (COMPAS), is a risk assessment algorithm powered by AI. COMPAS has been used in the United States to forecast which criminals are most likely to re-offend. However, it has been found to indirectly contain a strong racial bias, meaning black defendants are nearly twice as likely to be misclassified as higher risk than their white counterparts. This is not only unethical, but also unacceptable, for people to be disadvantaged in the application of these systems on a mass scale.
Transparency and Explainability
Another potential problem that comes with AI is the lack of transparency and explainability about what goes into the algorithms and how these algorithms work. The model Nvidia, didn’t look different from other autonomous cars, but it was unlike anything demonstrated by Google, Tesla, or General Motors. However, it showed the rising power of AI: the car didn’t follow a single instruction provided by an engineer or programmer. Instead, it relied entirely on an algorithm that had taught itself to drive by watching a human do it.
Following is a fairly good podcast that gives some overview of Nvidia’s self-driving cars.
Having such a self-driving car seems impressive.
However, the fact that no one indeed clearly knows how it made its driving decisions is a bit unsettling.
Self-driving car sensors capture and collect a tremendous amount of data every single second.
Such information about the environment has been feed straight into a huge network of artificial neural network model that process the data and then deliver the car operactions (control on the steering wheel, the brakes, and other systems) that required in react to the environment.
From the demo, the result seems to match the responses you’d expect from a human driver.
However, what if one day it did something unexpected—crashed into a tree, or sat at a green light?
Unfortunately, the system is often so complicated that it might be difficult to find out why.
Even the engineers who designed it may struggle to isolate the reason for any single action.
Accountability, Responsibility and Governance
Developments in autonomy and machine learning are rapidly enabling AI systems to decide and act without direct human control. However, there is also a need for debate around who will be accountable for decisions made by AI.
“For there to be trust in AI, there needs to be a degree of accountability”
– Ms. Kate Marshall, KPMG Law partner
Greater autonomy necessarily comes with greater responsibility, although the notion of responsibility might be significantly different between when it applies to humans and when it applies to machines. How can governance functions keep up with fast-paced development in AI? Who should be held accountable?
Take self-driving cars as an example, as stated by KPMG’s technology partner Kate Marshall, the current state of rules surrounding the testing of driverless cars and autonomous vehicles are too reactionary itself, “with no overall framework or agreement around the approach to artificial intelligence being used.”
Responsible AI is more than the ticking of some of the ethical ‘boxes’ or the development of some add-on safety features in AI systems. It requires the participation and commitment of all relevant stakeholders and the active inclusion of all of society. This means training, regulation, and awareness, etc. Watch the following discussion between Cathy Cobey, Partner in EY Canada’s IT Risk & Assurance practice and Dr. Cindy Gordon, Founder and CEO, SalesChoice discuss the issues in relation to Managing the Risks of AI.
1.5 Existing AI ethics and governance framework
AI is changing societies and economies around the world and the ethics of AI are of growing importance. In the last few years, countries worldwide (governments, research institutions, and communities, as well as industries, etc.) are developing solutions, principles, and guidelines for ethical AI (see Figure 1.1).
](figures/worldAIEthicsEffort.png)
Figure 1.1: Map of recent developments in artificial intelligence ethics worldwide Source from Artificial Intelligence: Australia’s Ethics Framework
In this section, we study a brief review of the existing governance framework, principles and guidelines to issues related to AI ethics around the world.
PRESCRIBED READING
Artificial Intelligence: Australia’s Ethics Framework: Chapter 2 Existing frameworks, principles and guidelines on AI ethics
DISCUSSION
- What impact is AI likely to have on Australian society in the near and longer-term?
- What are the ethical, trust and human rights challenges presented by AI?
- What are the AI Principles and standards that should be adopted for Australia?
- What does a governing body responsible for setting standards and guidelines for the ethical use of AI that would inform self-regulation and guide government regulation look like?
In this course, we are going to investigate each of these perspectives carefully, raise our awareness on the risks coming along with AI, and discuss potential solutions to handle them.