The ethical dilemma of AI
This article is authored by Mehdi Hussain, doctoral candidate, School of International Studies, Jawaharlal Nehru University.
Technological advancement in the form of artificial intelligence (AI) especially during the last three decades has brought a whole new meaning to human civilisation. It has been pushed by the progress in academic research and the growing appetite of big companies like Google, IBM, Amazon, Facebook and Netflix. Sophia, a humanoid robot developed by Hanson Robotics, brought excitement to the world with its launch in 2016. Sophia is celebrated as the world’s first robot citizen and the first robot Innovation Ambassador for the United Nations Development Programme. The social humanoid robot even appeared on the Tonight Show and Good Morning Britain and spoke at several conferences in different countries.
AI can serve as a “means” to human progress, according to the European Union’s High-Level Expert Group on AI. Human development is increasingly measured and studied using enormous data in research and development sector. However, individual progress and societal well-being for the common good based on AI innovation should be the “end” of AI and not in itself. AI can process huge data sets which man cannot process at the speed and accuracy it does. It is consistent and capable of performing repetitive tasks efficiently. It can find out patterns using data, for instance, commuting times or email spam filters.
Other ethical benefit of AI includes accelerating the path towards the UN SDGs with the help of AI global solutions. The International Telecommunication Union (ITU) organises the “AI for Good” Summit linking 40 UN agencies and co-convenes with the government of Switzerland. It brings in multiple stakeholders—public and private—to enhance collaboration efforts and equitable access to AI technologies. Through AI-enabled achievement of SDGs, it serves as an action-oriented platform for the UN; 2023 AI for Good Summit, thus, focuses on AI to advance health, climate, gender, inclusive prosperity, and sustainable infrastructure among others.
The existing technology gap between the countries of the global South and North creates a hindrance to digital equity. “AI for Good” platform can address to an extent the tension between North and South regarding the debate about switching to sustainable energy that often tarnishes the global climate change negotiations. The developing and least developed countries neither could afford the cost of the shift nor possess the technology required to achieve the sustainable development targets.
“AI for Good” promotes start-ups to grow their AI-powered and SDG-driven solutions through a pitching platform called AI for Good Innovation Factory. It also connects these start-ups to businesses, investors, governments and big IT companies. The World Food Program’s Hunger Monitoring Unit relies on machine learning to collect reliable food security data juxtaposing climate and economic data to boost the early warning systems in the case of extreme weather events. In this effort, AI Deep Learning can generate several possible climate models which are fed to policy-making.
Then in the entertainment media AI technologies have transformed through advances in storage and processing power, for instance, Netflix’s recommendation engine or Spotify’s automated song selection.
The negative effects of AI can be related to the technology-man relationship. it does bring out conflicting interests in the form of binaries such as freedom/lack of privacy, reliability/lack of transparency, inclusiveness/discrimination, profits/joblessness, efficiency/labour, democracy/power concentration, agency/dependency, justice/marginalisation, and so on. In this context, it is often debated about the acceptable cost or limitation of “living in a digital world”. It raises ethical concerns about the design, development and use of AI on humans in a technology-driven society.
AI’s inevitability in human progress also raises serious questions about the “balance” between these binaries. Automation of labour has created distress in the labour market. AI is replacing works ranging from simple and repetitive jobs to professional medical radiology and journalist reporting works.
ChatGPT—an AI chatbot released by OpenAI in November 2022—is a new technology mimicking human writings or voices through the use of language models to construct sentences or audio outputs for spontaneous conversations. On the one hand, AI robots can provide companions to fight the social isolation of adults. On the other, new AI can make man dependent on technology at the risk of diminishing human reasoning and, his isolation from society and himself. AI voice generator can generate voice cloning which can again violate ethical boundaries of privacy and consent. Thus, several lawsuits are common against such AI voice simulations. Bollywood actor Anil Kapoo won the legal battle against misuse of his name, image and voice restoring the principle of justice and fairness. The recent crisis-hit Hollywood over the AI takeover of creativity is an instance of real concern. AI-enabled content creation platforms can simulate a human actor, and write scripts or imagery. OpenAI was sued by the Game of Thrones author George RR Martin and other authors and novelists over the use of their books without permission for ChatGPT’s algorithm for language models.
Humanoid robots like MAiRA and 4NE-1 of NEURA Robotics fitted with advanced AI and unique sensory technologies allow human-robot interaction. MAiRA is available for commercial use. 4NE-1 can carry weights up to 20 kg and perform sensory functions like recognition of voices, languages and even voice tones and emotions. Though it can free man from dangerous and repetitive work, it can also create massive job displacement. Then, Digital Farmhand is a low-cost and solar-powered agro-tech AI robot to equip small farms to perform automated weeding, sampling, and crop manipulation. It can work for 4-15 hours to reduce the cost of labour and chemical input.
AI is significant in decision-making within big businesses through innovative solutions or classifications of predictions after layers of data processing. International Data Corporation (IDC) forecasts spending of AI systems will reach $154 billion in 2023 and touch $300 billion in 2026. Given these attractive AI-driven market opportunities, businesses need to adopt AI faster for fear of becoming redundant. It can develop business models that focus on competitive innovation in product designs based on improved supply and demand. Big companies use AI-based search engines as business models.
However, it can lead to the concentration of wealth in big companies. It is reported that Apple was valued at $2 trillion, and the combined stock market value of Apple, Microsoft, Amazon, Alphabet and Facebook increased by $3 trillion during the pandemic from March and August 2020. Concentration of power can also occur through the manipulation of elections. Electoral democracy can be influenced through algorithmic study of targeted electoral behaviour and influence them with appealing political messages, as in the Facebook and Cambridge Analytica scandal.
In the United States, self-driving cars have caused deaths and accidents or incidents of technical failures due to errors in software coding. For instance, several complaints were filed against Tesla for braking issues. Automotive AI uses the same principles of language models as ChatGPT which can be dangerous in comprehending unobserved factors of driving.
In this context, there should be mechanisms to check on AI systems when they do not perform as they should. For example, policy regulation of imperfect algorithms potentially risks erroneous medical diagnosis or wrong evaluation of the car environment. However, the lack of technical know-how across industry and government is another issue in this direction. The governments have to train their employees with technical competency in AI before regulating AI systems.
The AI for Good has called for an inclusive, safe and responsible AI to benefit countries with low technical capabilities. It reiterates on removal of biases to push for ethical and rights-based systems upholding the UN human rights and values. It thus appeals for transparency and accountability in AI systems. In future, proper deliberation among stakeholders should be made in time before any future possible social unrest, geopolitical instability, and economic disparity due to unchecked AI advancements. It, thus, serves as a motivating factor for countries to adhere to SDGs through regulation of AI systems and operations making sure it delivers digital equity and justice.
Nonetheless, the governance of algorithms is complex given the existing different government bodies or the need for new independent agencies. In addition to that, several layers of governance need to be in place to regulate AI algorithms at the levels of code or operators or manufacturers—government or private. Then, laws should take into consideration legal and ethical aspects of society keeping in mind the research and development norms. They should define user rights of privacy and transparent AI systems. Moreover, there is a need to focus on digital literacy for all. It has to go through trade-offs between the necessity of AI and its negative effects.
AI is arguably making inroads into every aspect of society. In a digital world, technology and man have become two sides of a coin. Its contribution to making the world a better place is significant with 280 projects focusing on AI capabilities to fight the challenges of climate change, improve education, eliminate hunger, and eliminate poverty and so on. While prevention and mitigation of AI-related harms is essential, generative AI is here to stay and its future is full of possibilities.
This article is authored by Mehdi Hussain, doctoral candidate, School of International Studies, Jawaharlal Nehru University.
