An Introduction
Artificial Intelligence (AI) is still a nascent technology with astronomical promises. Nevertheless, AI seems to have kicked up a storm of debates, with people passionately arguing in favor of and against AI. The fear of losing one’s identity, way of life, or livelihood seems to be motivators of resistance (1,2). From tractors to cell phones, all forms of innovation have suffered this rite of passage. Will AI have to submit to the same resistance?
While Stephen Hawking and Elon Musk warn about AI being an existential threat to humans, Mark Zuckerberg is optimistic about its phenomenal potential (3,4). Scrutinizing issues that lie within these extremes will allow us to explore the full range of the benefits and harms of AI. Could the past help us in this journey? What did the founding fathers of AI have in mind? A peek into the pages of AI history would help unearth their perspectives.
Tracing the Roots
According to Pamela McCorduck, an author who has written about the history and philosophical significance of AI, AI began with “an ancient wish to forge the gods” (5). In the modern times, the birth of AI can be attributed to the Dartmouth Summer Research Project on Artificial Intelligence (DSRPAI) in 1956, where attendees broke ground on the assumption that “every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it” (6). AI has been on an upswing since the late 1990’s thanks to three technological breakthroughs: affordable parallel computing, big data, and deep learning.
In the modern era, at the DSRPAI, automating the functions of the brain seemed to be the agenda (6). The goal of the attendees was to address issues like teaching computers the intricacies of language, rules of reasoning, conjecture, and manipulation of words. Among the problems they intended to analyze was simulating a network of neurons in order to model the brain and teach this model human skills like self-improvement, handling randomness, hunches, intuition and making educated guesses. They also wanted to work on removing computational barriers by improving the speed and capacity of the available computers (6).
The Good & The Bad
It is naive to assume that the applications of a new invention will only be virtuous. In fact, the flavors have always been good, bad and ugly. It takes a collaborative effort between scientists, businesses and the general public to translate an invention into successful products that serve the needs of society.
“Success in creating effective AI, could be the biggest event in the history of our civilization. Or the worst. We just don’t know. So we cannot know if we will be infinitely helped by AI, or ignored by it and side-lined, or conceivably destroyed by it,” said professor Stephen Hawking (4). This is a perfect example of the necessity of going through the good, bad and ugly.
Benefits to Society
Food, healthcare and personal safety. These form the core of the basic necessities for all human beings. AI’s success in these areas allows it to live up to its promises.
The Food and Agriculture Organization of the United Nations (FAO) predicts that by 2050, food production will increase by 50% (7). A growing population, shrinking arable land, and an aging and dwindling labor force contribute to this statistic. AI has helped farmers increase agricultural productivity.
Transfer learning is being used to teach AI to recognize diseases and pest damage in crops. AI can identify diseases with 98% accuracy, train cameras and sensors to capture images, identify weeds, spray the right herbicides, train robots to pick fruit, sample soil, analyze data to detect nutrient deficiencies, and undertake remedial measures (8). Microsoft’s FarmBeats is an example of such an AI-based agriculture platform (9).
Additionally, according to the National Institutes of Health (NIH), “precision medicine is an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle” (10). Google’s AI-based tool, DeepVariant, builds a more efficient personalized genome in a faster and cost-effective way (11). Companies such as Deep Genomics use deep learning to analyze personalized genomic data, identify patterns in that data that might contribute to diseases such as Alzheimer’s, cancer, heart disease, etc., and help drug companies with drug discovery (11).
The drug discovery process involves gigantic amounts of data, images, and research papers. It is slow due to the limitations of human researchers who can read only between 200 and 300 papers per year. With AI, such data can be processed by natural language processing algorithms, assimilated, correlated, and connected to related databases to expedite drug discovery. Moreover, the discovery of a drug component is extremely labyrinthine because it involves checking the numerous combinations between that component and other biological factors. AI aids in cutting down the time taken (12).
AI has also become an invaluable tool in protecting data. Algorithms using machine learning scan repositories of malicious programs identified from previous attacks, learn what to look for, and predict future attacks. AI is being used in the war against spam and phishing, to filter out violent images and illegal financial transactions, protect data, and detect computers infected with malware (13).
What do the Naysayers Think?
Naysayers are raising an alarm about AI creating a society of a few haves, who will own strong AI tools/skills/devices, and many have-nots. Since, unlike the previous revolutions, such as the industrial and computer revolutions, which replaced certain type of jobs with others, the AI revolution is ready to destroy numerous blue-collared jobs. The few left safe would share the profits and create a disparate number of have-nots (14, 15). Exploring how modern life will be harmed by AI is a worthy exercise.
As of 2015, the percentage of the population employed in agriculture in the world’s two most populous countries, China and India, is approximately 29% and 51% respectively (16). Even though AI seems to be taking roots in developed countries, it will not be long before it spreads to countries like India and China where it could disrupt the livelihood of major chunks of the population (17).
An often overlooked fact is that a majority of doctors are not trained to interpret results of AI-based precision medicine or relay these results to patients (18). This leads to misdiagnoses and mistreatments. Additionally, precision medicine needs access to the most personal data, one’s genome sequence. Therefore, it demands the utmost alertness because it touches issues of ethics and security. Who should be allowed access to this data? How is the privacy and security of this data ensured?
In the frenzy to import AI into precision medicine, the cost associated with it is overlooked. Mass adaptation of a product/service drastically reduces its price. Precision medicine, which aims to deliver individualized treatment, is therefore expensive (19). Are AI and precision medicine worsening the burden of soaring healthcare costs?
AI is also gaining an upper hand over humans in committing cybercrimes. Personally Identifiable Information (PII) is information that can be used to identify, contact, or locate a single person. AI-aided cybercriminals mine enormous amounts of data, extract PII, use it to steal identities, and enable hackers to mount personalized attacks (20). Phishing, considered to be a lower-level crime, has been empowered by AI. Experimental results showed that AI is remarkably superior to humans in distributing phishing messages over social media (21).
Hindering the livelihoods of large chunks of society and stealing private data could cause turmoil in societies that are unprepared to handle them. It is worth considering the opinions of the naysayers.
The Future
AI offers hope for providing cures for diseases such as Alzheimer’s, Parkinson’s, ALS, cancer, etc., increasing food production, making a commendable dent in hunger in poorer countries, and launching powerful defenses against cybercrimes. But is this a promise of utopia?
At the other end of the spectrum is the prediction of an AI-fueled dystopia: AI bots capturing PII, conquering human identities, overtaking human intelligence, starting wars, and causing high unemployment and societal unrest.
Having gone through the exercise of examining the good, bad and the ugly, it is prudent to ask the next logical questions. Should AI be regulated? What might be the issues within the purview of such regulations? Has the train already left the station?
Considering the speed at which AI is penetrating our lives, it is judicious to accept that AI cannot be turned back and work towards establishing a global consensus on AI regulation (22). It is wise to work to turn this powerful technology into an ally and prevent it from attaining complete autonomy.
Securing data, the lifeblood of AI, is paramount. The General Data Protection Regulation (GDPR), being proposed by the EU is gaining momentum worldwide. However, will it be universally accepted? Scrutinizing for bias in AI algorithms and deciding which AI decisions are ethical are some of the other critical issues to be regulated (23). Achieving a global consensus on these regulations expeditiously is advantageous to everyone.
Mythri is a first year in Matthews Hall
Works Cited
[1] Steven Overly. Washington Post https://www.washingtonpost.com/news/innovations/wp/2016/07/21/humans-once-opposed-coffee-and-refrigeration-heres-why-we-often-hate-new-stuff/?noredirect=on&utm_term=.316e1bbaef21 (accessed Oct. 2, 2018).
[2] Calestous Juma. World Economic Forum. https://www.weforum.org/agenda/2016/07/why-do-people-resist-new-technologies-history-has-answer/ (accessed Sept. 22, 2018).
[3] Camila Domonoske. NPR. https://www.npr.org/sections/thetwo-way/2017/07/17/537686649/elon-musk-warns-governors-artificial-intelligence-poses-existential-risk (accessed Sept. 20, 2018).
[4] Osborne, Hannah. Stephen Hawking AI Warning: Artificial Intelligence Could Destroy Civilization. Newsweek [Online], November 7, 2017. https://www.newsweek.com/stephen-hawking-artificial-intelligence-warning-destroy-civilization-703630 (accessed Sept. 18, 2018).
[5] McCorduck, P. Machines Who Think, 2nd ed; A.K. Peters, Ltd.: Massachusetts, 2004.
[6] McCarthy, John. et. al. AI Magazine. 2006, 27, 12-14.
[7] Food and Agriculture Organization of the United Nations. http://www.fao.org/news/story/en/item/471169/icode/ (accessed Sept. 19, 2018).
[8] Food and Agricultural Organization of the United Nations. http://www.fao.org/e-agriculture/news/can-artificial-intelligence-help-improve-agricultural-productivity (accessed Sept. 20, 2018).
[9] FarmBeats: AI & IoT for Agriculture. https://ai.intel.com/powering-precision-medicine-artificial-intelligence/ (accessed Sept. 19, 2018).
[10] Intel AI. https://ai.intel.com/powering-precision-medicine-artificial-intelligence/ (accessed Sept. 19, 2018).
[11] Knight, W. Google Has Released An AI Tool That Can Make Sense Of Your Genome. MIT Technology Review [Online]. 2017. https://www.technologyreview.com/s/609647/google-has-released-an-ai-tool-that-makes-sense-of-your-genome/ (accessed Sept. 20, 2018).
[12] Fan, P. How AI Can Speed Up Drug Discovery. AI Technology and Industry Review [Online]. https://medium.com/syncedreview/how-ai-can-speed-up-drug-discovery-3c7f01654625 (accessed Sept. 20, 2018).
[13] Lily Hay Newman. Wired. https://www.wired.com/story/ai-machine-learning-cybersecurity/ (accessed Sept. 18, 2018).
[14] Polonski, V. People Don’t Trust AI–Here’s How We Can Change That. Scientific American [Online], January 20, 2018, https://www.scientificamerican.com/article/people-dont-trust-ai-heres-how-we-can-change-that/ (accessed September 14, 2018).
[15] Darrell M. West. Brookings. https://www.brookings.edu/blog/techtank/2018/04/18/will-robots-and-ai-take-your-job-the-economic-and-political-consequences-of-automation/ (accessed Sept. 20, 2018).
[16] Max Roser. Our World in Data. https://ourworldindata.org/employment-in-agriculture (accessed Sept. 19, 2018).
[17] Evan Fraser and Sylvain Charlebois. The Guardian. https://www.theguardian.com/sustainable-business/2016/feb/18/automated-farming-food-security-rural-jobs-unemployment-technology (accessed Oct. 3, 2018).
[18] Advisory Board. https://www.advisory.com/daily-briefing/2015/02/10/the-challenges-for-personalized-medicine (accessed Sept. 19, 2018).
[19] Dan Mangan. CNBC. https://www.cnbc.com/2015/12/04/personalized-medicine-better-results-but-at-what-cost.html (accessed Sept. 21, 2018).
[20] George Dvorsky. Gizmodo. https://gizmodo.com/hackers-have-already-started-to-weaponize-artificial-in-1797688425 (accessed Sept. 21, 2018).
[21] Norton, S. Era of AI-Powered Cyberattacks Has Started. Wall Street Journal. https://blogs.wsj.com/cio/2017/11/15/artificial-intelligence-transforms-hacker-arsenal/ (accessed Sept. 22, 2018).
[22] Etzioni, O. How to Regulate Artificial Intelligence. The New York Times. https://www.nytimes.com/2017/09/01/opinion/artificial-intelligence-regulations-rules.html (accessed Sept. 23, 2018).
[23] Elizabeth Sablich. Brookings. https://www.brookings.edu/research/how-artificial-intelligence-is-transforming-the-world/#_edn3 (accessed Sept. 24, 2018).
Harvard Science Review 