Artificial Intelligence: Boon or Curse? by Prachi Saswade

Introduction

Artificial Intelligence (AI) is a term that has been floating around for a couple of decades. The definition of AI has been evolving, but the most widely accepted was given by John McCarthy in 2004, in his paper, What is Artificial Intelligence?: “It is the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable”.

The definition of AI has multiple approaches, but remain in line with the word, “intelligence”, specifically intelligence that is akin to that of humans. However, in the widely renowned authority in the field, Stuart Russell’s textbook, Artificial Intelligence: A Modern Approach, states that, "AI is concerned mainly with rational action. An ideal intelligent agent takes the best possible action in a situation” (Russell & Norvig, 2021).

The multiple mentions of “intelligence" and “rationality” date back to Alan Turing, who in his 1950 paper, Computing Machinery and Intelligence, asked the question, “Can machines think?”. This was succeeded by the infamous “Turing test”. This paper started the conversation about AI, and while the test has been controversial over the years, it is an important part of the AI history.

Shortly after Turing’s paper, John McCarthy coined the term “Artificial Intelligence” during the first AI conference held at the Dartmouth College, New Hampshire, United States. In the same year, Allen Newell, J.C. Shaw, and Herbert Simon, created the first ever running AI program, called the Logic Theorist. About a decade later, Frank Rosenblatt built the Mark 1 Perceptron, a computer based on neural networks that learn through “trial and error”.

The 1980s saw a rise in the usage of the back-propagation algorithm that allowed the neural network to train itself. These networks were then used in AI applications. Soon after, in a historical feat by IBM, “IBM's Deep Blue beats then world chess champion Garry Kasparov, in a chess match (and rematch)” (IBM Education, 2020).

AI has evolved through multiple trials, and based on the concept of applying human-like intelligence and rationality to computer decision making. The many industries working in the field have seen the massive adoption of the technology and continue their research on making it more and more “human” each day.

Level of Involvement of AI in our Daily Life

Artificial Intelligence has crawled into our lives and has become an integral part of it. AI is used in multiple fields in the present day that many-a- times we are not even aware that AI is being used there. Currently AI is being used all around the globe day and night. Forbes created a list of ten examples of how AI is used in its article, The 10 Best Examples Of How AI Is Already Used In Our Everyday Life. The list includes, technologies like FaceID, social media, digital voice assistants, etc. Starting with the most basic one, unlocking your phone. Apple’s FaceID technology uses artificial intelligence and 3D scanning to register the user’s face demographics. “It then uses machine learning algorithms to compare the scan of your face with what it has stored about your face to determine if the person trying to unlock the phone is you or not” (Marr, 2019).

Continuing on the same spectrum, social media uses AI to curate each user’s feed based on their history of liked posts, and engagement to certain content. The machine learning algorithms also aid in filtering out false news and content that multiple users have engaged with. Engagement with content requires the accompanying text to be written well, this requires tools such as Grammarly. Grammarly uses AI and natural language processing to ensure that its users focus on writing and leave the grammar to Grammarly. This technology is being used for professional emails, and any other writing that a user might require.

Many smart homes are in trend now and the the vital component of these smart homes are smart home devices like Alexa, Google Assistant, and HomePod. These smart devices use AI to to keep learning from the user’s usage patterns. Amazon recommendations is one of the more well known AI technology. Based on the user’s past order history, and searches, the recommendations for products are personalised. Along the same lines, Netflix uses AI and the past viewing history in the same way. Netflix is known for its spot-on personalised TV show and films recommendations for its consumers.

Lastly, “Google maps and other travel apps use AI to monitor traffic to give you real-time traffic and weather conditions as well as suggest ways to avoid gridlock” (Marr, 2019). Many car companies now have an in-built mapping system in the cars and this further allows the user to commute with ease. The AI technology is only growing more and more each day, and is being integrated into our lives rapidly. It is only time before every commodity we use will have some enhancements made to it to accommodate artificial intelligence.

Various Approaches to integrate AI seamlessly

Approaches to AI has a different connotation in the setting that it is being used in. For example, approaches to that drive AI research includes – cybernetics, symbolic and sub-symbolic approaches, as well as, the statistical approach (Milošević, 2013). At the same time when one talks about integrating AI into business, the approaches change from concrete terms to an instruction manual, almost all ending with the advice to “start small”. In the general context, however, we have four main approaches to AI – reactive machines, limited memory, theory of mind, and self-awareness. These four approaches are based on the behaviour of the machines that will use AI.

1. Reactive Machines The most basic AI systems are based on reactivity only. These machines often are good at predictions based on a certain set of rules, games such as chess. These machines only “react” to a situation, with no meaning of the past. They have no memory of the past, and only works in the present moment. IBM’s Deep Blue, the chess-playing computer is a notable example of this approach. “Deep Blue can identify the pieces on a chess board and know how each moves. It can make predictions about what moves might be next for it and its opponent. And it can choose the most optimal moves from among the possibilities” (Hintze, 2016). This means, Deep Blue only processes the chess pieces in front of it at present and chooses its next move. It does not look back for any previous references. AI researcher Rodney Brooks, in his paper argues that all machines should be built on this system. His reasoning for this was that the programming for such stimulated worlds was often not accurate enough and did not provide a valuable “representation” of the world (Brooks, 1991). Reactive machines can be “easily fooled” because they have no concept of the world outside of the rules they are set within. These machines however, can prove to be extremely impartial as they only react to what is presented to them in real-time. This suggests that might prove to trustworthy due to lack of emotional engagement.

2. Limited Memory The limited memory machines are considered the Type II class machines. These machines have an ability to look into the past. The best example of this is seen in self-driving cars. Self-driving cars require the programmed world to have representations that are pre-programmed, such as traffic rules, or routes in the city, etc. These are also included when the car has to change its lane and avoid an accident. “But these simple pieces of information about the past are only transient. They aren’t saved as part of the car’s library of experience it can learn from, the way human drivers compile experience over years behind the wheel” (Hintze, 2016). It has been noted by both Brooks, and Hintze that it is difficult to build AI systems that are full of representations, as well as, remember experiences and learn how to tackle newer situations. Hintze has applied the Darwinian evolution to his research to let machines build their own representations.

3. Theory of Mind “[Theory of mind is] skill that involves the ability to think about mental states, both your own and those of others” (Cherry, 2021). This psychological concept introduces the next class of machines. These machines are far more advanced and as the theory of mind suggests, form representations of not only the world, but also about other participants or agents that exist within it. The example of this would be Sophia, the AI robot. As one see, Sophia can not only answer questions, but connect to various entities around her.

4. Self-awareness Self-awareness is the last approach to AI systems. This system is the most advances class of machines, wherein, the machines can build representations about themselves. Many researchers are looking to build AI systems that have a consciousness, not just understand it. This is a step up from the theory of mind, as here, the machines will be able to make inferences about other entities in the same way human rational thinking does. According to Hintze, “we are probably far from creating machines that are self-aware, we should focus our efforts toward understanding memory, learning and the ability to base decisions on past experiences” (Hintze, 2016). This however, does not dampen the possibility that we might live in a world where AI systems will be advanced enough to have a consciousness.
   

Impact of AI and its Major Benefits

Artificial Intelligence is used in almost every sector in the world today, extensively in the business world. There are many discussions about the impact of AI, both positive and negative. The impact of AI has brought on multiple questions, especially ones around employment of labour.

In his paper, The Forthcoming Artificial Intelligence (AI) Revolution: Its Impact on Society and Firms, Spyros Makridakis discusses the impact of AI on developing countries. According to him, this revolution will be more “pronounced” in the developing states for two reasons. Firstly, the use of machinery will replace the skilled and unskilled labour, this will result in foreign (developed) countries to remove their investments in the still developing countries. Secondly, “developing countries will be at a disadvantage by not being able to invest in expensive AI technologies, particularly since such technologies will reduce the demand for human labour thus further increasing unemployment” (Makridakis, 2017).

To solve this, Makridakis suggests that “[educating] their young people in AI technologies and by doing so become able to attract investments from abroad as well as manage to exploit the “sharing economy” (Makridakis, 2017). However, he also emotions that his might prove to be very difficult. The impact of AI will soon be seen in almost every factory across the globe, but in order for everyone to adopt the technology, the acceptance for it must be present.

However, all of AI is not bad news. There are multiple benefits to AI technology. For starters, it helps for smarter business decisions. It can also help, in enhancing the customer experience, medical advances, research and data analysis, solving complex problems, among others ("Top 10 Benefits of Artificial Intelligence (AI) | 10xDS", 2020). AI is also great at minimising errors and completing repeated tasks. This is extremely beneficial for companies that use data mining for decision making, and other activities. One example of this would be the clickstream analytics. This technology is used by multiple social media apps, as well as, companies like Amazon. They use the data generated when the user visits the website of an advertised product or service. This data then uses AI to target similar ads to the consumer, which for companies like Amazon, leads to the consumer purchases. Another benefit seen in this field is use of AI in chatbots. Chat bots are present on almost every company’s website today, and these are often run by AI. The AI scans through the frequently asked questions to provide an answer to the user within seconds. This technology reduces the time and allows a filtration of questions sent to the (human) employee to answer. These chat bits are now being used by banks as well. This technology is evolving rapidly and steadily. The integration of AI into our lives is increasing by the day, and like most technology invented to date, will only serve to make our lives easier. That being said, one must not ignore the problems that it comes with.

Associated Problems and Pitfalls AI has been a game changer in many sectors of the world. However, there have been many negatives attached to the technology. As mentioned before, one of the associated pitfalls is the impact the technology has on the developing countries. Other than that, there are multiple common challenges in AI such as, computational or hardware problems, lack of trust, lack of human-level experimental management, data security and privacy issues, and lastly the biases in the dataset.

As the world moves on to work with AI, the hardware for such upgradation requires enough cores and GPUs to work efficiently. This can take a monetary toll on any small company that is just starting up. Moreover, any company that is planning to move to AI will have to consider their options and make financially beneficial decisions. The lack of trust stems from the unknown networks that deep learning uses to come to conclusions. The logic is still muddy and can cause a string of worry for the users. This also brings to light the “human experience” into play. Humans use experiential knowledge to make further decisions. While one can argue that AI does the same, human accuracy based on other factors (social, economic, and cultural) is far greater.

Data privacy and security have been in the spotlight, especially since the FaceBook privacy case. The data that deep and machine aligning models use comes from across the globe and is generated by a large volume of users. The company collecting the data needs to be trustworthy, and it goes without saying that many companies might not always have good intentions with their clientele’s data.

Lastly, the issues of biased dataset. Unfortunately, a large portion of the data that the algorithms receive is biased. The bags may be based on religion, gender, or race. The data collected can also be biased in the way the algorithm is programmed, i.e.. the programmer or interpreter’s biases can come into play in these situations. These issues can seem daunting, especially for those who are new to this territory. However, AI algorithms can be created to reduce biases, and for this reason AI ethics exists. These ethical guidelines are followed around the world and reduce the negatives in this technology.

Proposed Applications of AI in Coming Years

AI has shaped the tech world, and given it a new form. According to IBM, AI advances would not be possible without a formula that contains three things: “the rise of big data combined with the emergence of powerful graphics processing units (GPUs) for complex computations and the re- emergence of a decades-old AI computation model—deep learning” ("The new innovation equation", n.d.). The future of AI will see these elements have a makeover. The rise of small data, and deep reasoning will be seen soon. According to the University of Southern California’s researchers, AI will change the entertainment industry, medicines, cybersecurity, vital tasks like help for the elderly, and transportation (Gammon, 2017).

Netflix has been using AI and machine learning techniques for a while now, and it will only get better. The addition of more streaming platforms can revolutionise the entertainment industry in the near future. With the help of AI, “film studios may have a future without flops: Sophisticated predictive programs will analyze a film script’s storyline and forecast its box office potential” (Gammon, 2017). Additionally, a user can also ask these platforms to create “virtual actors” and make a custom movie right at home.

A more personalised approach to medicine can be seen on the horizon. With genome sequencing technology already in boom, the medicines that a patient might need can be altered to the patient’s genome and provide for a more effective treatment. Moreover, AI will help health care analyse a patient’s health based on more factors like lifestyle, environment, and genes. The detection of any tumours, or diagnosing basic ailments will also be done by AI. Having a large volume of data generated by users of a certain application comes with the potential risk of hackers and data breaches. “There were about 707 million cybersecurity breaches in 2015, and 554 million in the first half of 2016 alone” (Gammon, 2017). According to USC, AI’s ability to self-learn and automate can be a fruitful solution to remain one step ahead of the hackers. This will ensure the security of billions of people across the globe.

Security and safety are utmost important human values, but so is independence. Many elderly citizens around the globe struggle to do daily tasks, or often require someone keeping an eye on them. With the working culture, they are usually left to look after themselves. AI tools around their areas of living can provide for a monitor on their movement, as well as, help with reaching objects on a high shelf, and ensure the supply of nutritious food. Moreover, these tools could mow their lawns, and help with maintaining the general hygiene of their residence. Additionally, AI assistance can be extremely useful in tasks such as mining, firefighting, and handling dangerous materials. We are already seeing a rise in self-driven cars. However, in the future this might expand into the public transportation systems as well. These AI driven vehicles are often much safer than humans, as they never get distracted but he radio or the other passengers in the cars. These are just the proposed application of AI, and there definitely will be more as the days pass by. The importance of AI will just increase multi-fold and defining only a certain amount plausibilities of its future can prove to be limiting its true potential.

Future Predictions – Boon or Curse

AI has seen a slow burn for a while but is deemed to explode into every aspect of our lives soon. That being said, the question still remains, is Artificial Intelligence a boon or a curse? AI has more benefits than we can count, and like every technology ever invented, it is here to make our lives easier and better. AI has seen better healthcare, better production, and better decision making. One cannot argue that AI saves us from repetitive and ‘boring’ tasks form time to time. Additionally, its capacity to sift through large volumes of data, or big data, is unmatched. To repeat the same tasks but using only human workforce will take years.

That being said, AI also comes with its own pitfalls. Relying on technology can make some people wary, especially with multiple security and data privacy issues. According to multiple people, AI still does not understand human values like privacy, and in many ways cannot match a human’s emotional and social intelligence. AI can only use the provided information and come to conclusions based on the algorithms provided by the programmer, and is quite redundant by itself, unlike humans.

AI can only be more “like” humans, but cannot be completely “humane”. AI when looked at as a tool can provide for millions of possibilities, and that might be the best way to look at it. AI can be used for multiple mad- practices, and ethics can only get one so far. Ethics are important, and in order for every user of AI to implement and respect them, there need to be strict judicial laws across the globe to ensure the safety of the people.

The technology is still evolving, and it might be wise to wait a little longer to categorise it as a “boon” or a “curse”. No technology can ever fit into only one category, each one comes with its own pros and cons. With AI, we might need to wait until we can see which one outweighs the other.

Bibliography 1. McCarthy, J. (2007). What is artificial intelligence?. 2. Russell, S., & Norvig, P. (2021). Artificial intelligence: A Modern Approach (4th ed.). Pearson Education Limited. 3. Turing, A. M. (2009). Computing machinery and intelligence. In Parsing the turing test (pp. 23-65). Springer, Dordrecht. 4. IBM Education. (2020). What is Artificial Intelligence (AI)?. Ibm.com. Retrieved 28 June 2022, from https://www.ibm.com/cloud/learn/what-is- artificial-intelligence. 5. Marr, B. (2019). The 10 Best Examples Of How AI Is Already Used In Our Everyday Life. Forbes. Retrieved 28 June 2022, from https:// www.forbes.com/sites/bernardmarr/2019/12/16/the-10-best-examples-of- how-ai-is-already-used-in-our-everyday-life/?sh=205c28bf1171. 6. Milošević, N. (2013). Approaches to artificial intelligence. Inspiratron.org - Natural language processing, machine learning and cybersecurity. Retrieved 1 July 2022, from https://inspiratron.org/blog/ 2013/05/10/approaches-to-artificial-intelligence/. 7. Hintze, A. (2016). Understanding the Four Types of Artificial Intelligence. GovTech. Retrieved 1 July 2022, from https:// www.govtech.com/computing/understanding-the-four-types-of-artificial- intelligence.html. 8. Brooks, R. (1991). Intelligence without representation. Artificial Intelligence, 47(1-3), 139-159. https://doi.org/10.1016/0004-3702(91)90053- m. 9. Cherry, K. (2021). Why the Theory of Mind Is Important for Social Relationships. Verywell Mind. Retrieved 1 July 2022, from https:// www.verywellmind.com/theory-of-mind-4176826. Artificial Intelligence (AI) – Boon or Curse? 17

10. Makridakis, S. (2017). The forthcoming Artificial Intelligence (AI) revolution: Its impact on society and firms. Futures, 90, 46-60. https:// doi.org/10.1016/j.futures.2017.03.006 11. Top 10 Benefits of Artificial Intelligence (AI) | 10xDS. 10xds.com. (2020). Retrieved 2 July 2022, from https://10xds.com/blog/benefits-of- artificial-intelligence-ai/. 12. Vadapalli, P. (2021). Top 7 Challenges in Artificial Intelligence in 2022 | upGrad blog. upGrad blog. Retrieved 5 July 2022, from https:// www.upgrad.com/blog/top-challenges-in-artificial-intelligence/. 13. The new innovation equation. IBM Cognitive - What's next for AI. Retrieved 7 July 2022, from https://www.ibm.com/watson/advantage- reports/future-of-artificial-intelligence/ai-innovation-equation.html. 14. Gammon, K. (2017). 5 Ways Artificial Intelligence Will Change the World by 2050. USC News. Retrieved 7 July 2022, from https:// news.usc.edu/trojan-family/five-ways-ai-will-change-the-world-by-2050/.