Artificial Intelligence (AI) is a field of computer science that encompasses a myriad of disciplines focused on improving computer programs to become self-learning. From natural language processing (NLP) to image classification to sound analysys, AI programs are “trained” on large training datasets allowing them to recognize patterns and act accordingly. Once an AI program is trained, it can be used to analyze real world data to deliver analysis and decisions. AI programs can also be trained against other AI programs using a technique called generative adversarial networks (GAN) wherein one AI program tries to learn while an adversarial AI program tries to defeat the first one. This way both AI programs train against each other and get better.
Most of the distruptive power of AI comes from disintermediation of humans. This happens by doing something that humans can’t do, usually because of complexity or scale. This allows rapid detection of fraud, crowd safety, identification and access, disease detection, advance warning of disease outbreaks, and just about any field involving typically laborious analysis of large amounts of data. Because of the speed of analysis, AI is also the driving force behind autonomous navigation industry such as self-driving cars, drones and underwater transport, thereby disrupting typically human-driven transportation and delivery systems.
The following are a few of the many fields where AI is making a differnce already:
Telemedicine allows health care professionals to evaluate, diagnose and treat patients at a distance using telecommunications technology. It permits two-way, real time interactive communication between the patient, and the physician or practitioner at the distant site.
Image analysis is being used for training classification software on biopsy scans
Visual recognition (especially on the health, satellite and smart cities tracks) and image recognition/ classification;
Algorithmic analysis of (e.g.) satellite imagery or Amazon and Netflix recommendations;
Voice recognition and intelligent captioning systems;
Fraud detection and other financial services by credit card companies.
Text recognition and analysis, natural language processing and spontaneous conversation;
Increasingly robots and other automated systems will use AI for manual or cognitive tasks.
The behavior of humans is completely opposite from that of machines. Humans behave like humans – forgetful, error-prone, variant, delightful, irritating. Machines have none of those qualities except, perhaps irritating. From self-driving cars on public roads to self-piloting reusable rockets landing on self-sailing ships, machine intelligence is supporting or entirely taking over ever more complex human activities at an ever increasing pace. The greater autonomy given to machine intelligence in these roles can result in situations where they have to make autonomous choices involving everything from copyright violations to who lives or dies. This calls for not just a clearer understanding of how humans make such choices, but also a clearer understanding of how humans perceive machine intelligence making such choices. Recent scientific studies on machine ethics have raised awareness about the topic in the media and public discourse.
With the rising use of AI in health, there is a concern about “doctor disintermediation.” With the possibility of a wrong diagnosis or a missed diagnosis slipping through the system, there will always be the need for human supervision. Calibration of the balance between automation and human-intervention will have to be continuous and responsive to errors.
Similarly, in the domain of self-driving cars, there is the dilemma of choosing the lesser of two bad options. Experiments such as The Moral Machine (see Resources below) are trying to determine how the algorithm for doing so will have to be programmed, for this will be a function of our prevailing cultural and legal norms.
Just as AI can be used to mimic real sources to create fake content, AI can also be deployed to detect fake content. There is a real concern about the generation of fake news, both by human and algorithmic sources. Apple has chosen to use human curation along with algorithmic newsfeed generation to counter fake or misleading news.
The 7 Steps of Machine Learning (AI Adventures). https://www.youtube.com/watch?v=nKW8Ndu7Mjw
Introduction to Machine Learning. https://www.youtube.com/watch?v=h0e2HAPTGF4
Moral Machine – Human Perspectives on Machine Ethics. https://www.youtube.com/watch?v=XCO8ET66xE4
ITU AI Repository. https://www.itu.int/en/ITU-T/AI/Pages/ai-repository.aspx
AI for Good Summit 2018 Report. https://2ja3zj1n4vsz2sq9zh82y3wi-wpengine.netdna-ssl.com/wp-content/uploads/2018/12/SDGs-Report.pdf
Hamilton and Amabel James Center for Artificial Intelligence and Human Health. https://www.mountsinai.org/about/newsroom/2019/icahn-school-of-medicine-at-mount-sinai-to-establish-world-class-center-for-artificial-intelligence-hamilton-and-amabel-james-center-for-artificial-intelligence-and-human-health Paige. https://paige.ai
Artificial Intelligence and Life in 2030. https://ai100.stanford.edu/sites/g/files/sbiybj9861/f/ai_100_report_0831fnl.pdf
Whom should self-driving cars protect in an accident. https://www.progrss.com/movement/20180216/trolley-problem/"
A Voting-Based System for Ethical Decision Making. https://arxiv.org/pdf/1709.06692.pdf
Discriminating Systems: Gender, Race and Power in AI. AI Now Institute. https://ainowinstitute.org/discriminatingsystems.html
Youtubers and Record Labels Are Fighting, and Record Labels Keep Winning: The battle over copyright continues. https://www.theverge.com/2019/5/24/18635904/copyright-youtube-creators-dmca-takedown-fair-use-music-cover
We’ve Been Warned About AI And Music for Over 50 Years, But No One’s Prepared. https://www.theverge.com/2019/4/17/18299563/ai-algorithm-music-law-copyright-human