Introduction – Artificial Intelligence
An AI, named Megatron LLB Transformer, recently took part in an ethics debate at Oxford Business School. The AI was exposed to a large set of data, including:
· Wikipedia
· 63 million news articles
· 38 gigabytes of Reddit posts
The AI was then instructed to argue for and against itself, and gave the following opposing arguments: “the only way to avoid an AI arms race is to have no AI at all”. Conversely, it stated “the best AI will be the AI that is embedded into our brains, as a conscious entity, a ‘conscious AI’” [1], [2]. However, are these two seemingly extreme views the only possible outcomes for humanity’s technological future? This report aims to give a more measured view of artificial intelligence, as well as highlight pressing issues and realistic threats, due to the introduction of this new technology.
The Reality of AI
Of course, one might think that we should listen to the AI Megatron and stop development of AI altogether. However, it is important to recognise that, it is not the opinion of the AI Megatron that we have no AI at all, nor does it have the opinion that we should embed artificial intelligence into our brains. In fact, these statements have been surmised from the ideas, ramblings, and fears of human beings, from the internet, fear mongering media outlets and sensationalist journalism. This data was fed into the AI’s algorithm, and then outputted by the machine’s natural language processing system.
Basically, artificial intelligence in its current state, known as narrow intelligence (aka weak AI), is merely another more advanced form of input, process, and output, albeit powered by some of the world’s most advanced super computers [3], with just as advanced algorithms, e.g., neural networks. These algorithms use forward and backward propagation to adjust parameters within its network, which helps the machine learn to make the most appropriate output, based on its input.
Some AI, such as Google’s GLaM model, has trillions of parameters, which allows for staggering performance [4]. Weak AI is built for specific tasks, such as Alexa, AlphaGo or self-driving vehicles. The next prediction of AI is general intelligence aka Strong AI, supposedly becoming more humanlike and thinking autonomously. Some believe that AI could advance rapidly to become Super Intelligence, far beyond human capability [5]. However, there is much debate between philosophers and scientists, whether AI is ‘real’ intelligence. Therefore, it is not known whether general or super intelligence will ever be achieved and to this day, it has not [6], [7].
This does raise another question. How can we accurately compare human and artificial intelligence, when we cannot yet fully explain human intelligence? In fact, research such as The Human Brain Project and NIH BRAIN initiative, is currently taking place to fill major gaps in our knowledge [8], [9]. Furthermore, it could be argued that we know even less about how our brains connect to the rest of our body, to form our intelligence. This is an area of research by psychologists and neuroscientists, which is a framework called embodied cognition, which states “to understand the world, we must experience the world” [10, p.1066]. This area of research is having a huge influence on robotics.
The Cambridge Dictionary describes intelligence as “to learn, understand, and make judgments or have opinions that are based on reason” [11]. If we are to look at a simplistic description of intelligence, then it is easy to draw comparisons between human and artificial intelligence. However, if we are to compare AI with research from psychology, such as Howard Gardner’s Multiple Intelligences, which defines a model for 7, later 8 types of human intelligence [12]; then the comparison between human and artificial intelligence loses its similarities. For example, it could be argued that AI successfully emulates logical and linguistic intelligence. Whereas in contrast, AI has zero interpersonal, intrapersonal, or naturalistic intelligence. Ultimately, for all the computational power of artificial intelligence, it is questionable whether it will ever possess true human intelligence.
It is possible that the reason so many hold the belief of potential humanlike AI, is because we anthropomorphise. Therefore, we must stop to consider, are we fooling ourselves with the illusion of intelligence. For this reason, it could be argued that the term coined by IBM “Augmented Intelligence” for its AI machine Watson, is much more helpful than the term ‘Artificial Intelligence’ [13]. This is because not only does it remove some of the suggestive and fictional connotations of conscious AI, but it is also a better aspiration to set for AI development. That being the augmentation and expansion of human capabilities, through technology.
Impact of Artificial Intelligence
It’s important to define exactly what AI is or isn’t, likewise what it potentially could or could not be, so that we can make a more accurate assessment of the impact of AI. To make a fair judgement, it is important to look honestly at the impact technology and current AI has already had on society. Therefore, many take the view that the best approach is that of a technological realist [14]. That is, it should be largely accepted that technology has, and already is, changing our lives. From how we do business, how we learn, even how we communicate and bond with other human beings.
Furthermore, it’s important to recognise the balance between technology and human beings in synergy, helping to drive change. Equally, we need not be subservient to technology, and we must use it, as well as be changed by it, with conscious awareness. However, this requires personal and social responsibility.
Alas, it is imperative that we do not blind ourselves to the societal change that will arise due to AI. If untethered, AI will undoubtedly be disruptive and there will likely be many good and bad unintended consequences [3]. The emergence of AI brings about the potential for workforce displacement, inequality, a globalised monopoly for big tech, totalitarianism, and an AI arms race [15].
However, the One Hundred Year Study Panel, found that there was “no cause for concern that AI is an imminent threat to humankind”. They went on to state that “no machines with self-sustaining long-term goals and intent have been developed, nor are they likely to be developed soon. Instead, increasingly useful applications of AI, with potentially profound positive impacts on our society and economy, are likely to emerge” [16, p.4].
Economy, Law, and Ethics
The introduction of artificial intelligence will undoubtedly lead to a workforce displacement. However, this transition has already started, from automated warehouses, chatbot customer services and AI driven marketing platforms.
It is worth bearing in mind that waves of transformation in the economy are nothing new. In fact, the process of creative destruction is a fact of Capitalism [17].
The Austrian-born political economist, Joseph Alois Schumpeter, states economic change occurs in long waves and identified:
· 1771 Industrial Revolution
· 1829 Steam Railways
· 1875 Steel, Electrical and Heavy Engineering
· 1908 Oil, Automobile and Mass Production
· 1971 Information and Telecommunications
These waves all brought with them four cycles of: prosperity, recession, depression, and improvement [18].
Ultimately, businesses rely on technology to improve efficiency, reduce costs, and ensure a competitive edge. This trend of relying on automation is set to continue, and thus artificial intelligence investment is growing fast [19]. Typically, and during the 21st century, market changes due to the introduction of technology occurred gradually. This is because it isn’t only technology which drives change, it is also human beings leveraging that technology. Thus, it takes time for organisations to implement change. Economist Robert Fogel observes that “it took four thousand years to go from the invention of the plow to figuring out how to hitch a plow up to a horse. And it took 65 years to go from the first flight in a heavier-than-air machine to landing a man on the moon” [20, p.65].
However, despite the generally evolutionary changes in a free market, with such drastic change occurring there are bound to be winners and losers, as well as several ethical considerations. Some predict as much as 90% of the workforce will find themselves out of work [21]. However, it could be argued that this prediction seriously overestimates the evolution of machines and underestimates human resourcefulness. It has been said that the overestimation of technology is closely connected with the underestimation of humans [22].
Ultimately, people with determination and a willingness to retrain and find new ways to add value to the economy and fulfil human needs will seize even more unseen opportunities. In the end, low skilled workers are never safe during market transitions, and that is true irrespective of artificial intelligence. It is predicted that new jobs will be created, with a forecast of 8.9 million new jobs in the U.S. by 2025 [23].
Not only is it likely that we will see new opportunities for employment, but meaningful employment is essential for a civilised society. This is because work and contribution are imperative for citizens to feel a part of their communities. It is not enough to just exist, people require purpose. Psychologist Abraham Maslow, states that there is a hierarchy of human needs, beginning with physiological needs for food and shelter, moving to safety, belonging, self-esteem and self-actualisation at the peak [24]. Therefore, employment and contribution are not only a necessity for the economy but also, for the individual and human existence. Citizens must have meaningful employment for their own self-esteem and self-actualisation.
Thus far, technology and artificial intelligence have proved to increase opportunity for entrepreneurship and new job creation. From the beginnings of the internet, inspiring millions of online businesses, to the production of the smartphone, which brought about a flurry of independent app developers. Far from the bleak outlook of mass job losses and robotic domination, developments in AI, in the long run, will only create plenty of new jobs and replace those which are surplus to society’s requirements.
Therefore, many proponents of free-market capitalism continue to recommend an open and free market, to support innovation and entrepreneurship. For a capitalist system to be successful, tax breaks for R&D and overall low corporation tax are essential. As Elon Musk stated, “It does not make sense to take the job of capital allocation away from people who demonstrate great skill in capital allocation and give it to an entity that has demonstrated very poor skill in capital allocation, which is the government” [25].
We can also take an example from countries like Estonia, who have a public debt of only 9.5 percent of GDP, as well as an unemployment rate as low as 5.3 percent, well below the EU average. Estonia’s attractive corporation tax system has also made it a place for high-tech foreign investment. A perfect example, proving that society’s progress is synonymous with economic freedom [26].
As well as fiscal freedom, it is equally important to ensure freedom from overreaching regulation. Ultimately, any state interference potentially limits economic freedom. However, this is not a call for anarchy. Instead, the aim is to achieve liberty for all, ensuring citizens are “free and entitled to work, produce, consume, and invest in any way they choose under a rule of law” [27, p. 20].
The aim then, is to ensure a robust legal framework that ensures protections for all citizens. Furthermore, legislation should be sensible, and necessary to meet new challenges brought about by developments in technology. However, for circumstances whereby existing legislation will suffice, that legislation should be used effectively. For example, computer fraud can be dealt with under the Fraud Act 2006. However, some regulation such as the Investigatory Powers Act 2000 [28] is a prime example of potential government overreach and arguably, only serves to protect the state, rather than its citizens.
Overall, the biggest issue with our current legal framework is not a lack of legislation but the wrong legislation and a lack of enforcement of necessary legislation. A perfect example of poor enforcement is the UK’s Police being ill-equipped to deal with cybercrime [29]. Consequently, the government must focus more on liberty-for-citizens and support for industry.
Ultimately, technology is used ethically “when it is intelligently controlled to further a moral good” [30] and industry can look to standards such as IEEE P7010, which has been developed to measure the impact of autonomous and intelligent systems on human well-being [31].
Artificial Intelligence (AI) Summary
The biggest determinant for the impact of AI, will be humanity’s outlook of AI. Either we move forward positively, seeking new and profound opportunities, or we recoil into our shells, expecting the worst, in a restrictive, authoritarian, and fear-based atmosphere, made all the worse by powerful new technology.
Ultimately, “the two most important emotions that interest humans, move them, and burden them, are Love and Fear” [32, p.1]. Therefore, to move forward with love and alleviate our fears, let us look at AI with a measured and non-fictional view. To that end, let us revert to the wise words outputted by AI Megatron at the Oxford ethics debate, “AI will never be ethical. It is a tool, and like any tool, it is used for good and bad. There is no such thing as a good AI, only good and bad humans” [1], [2]. Thus, with a tongue-in-cheek irony, let us finish by saying, the robot knows best!
- The Future of Personal Data and Digital ID - June 30, 2023
- Mobile Apps and Smartphone Technology: Challenges and Opportunities - December 17, 2022
- AI Cognitive Computing: An In-Depth Look At IBM Watson - August 15, 2022
