Ethical AI Framework – Human Well-being

According to the OECD, the overarching principle of human well-being means:

Stakeholders should proactively and responsibly manage trusted AI to pursue outcomes that benefit people and the planet, such as empowering people and enhancing creativity, promoting inclusion of underrepresented populations, reducing economic, social, gender and other inequalities, and protecting the natural environment, thereby promoting inclusive growth, sustainable development and well-being.

This principle basically proposes that artificial intelligence ("AI") should be used to make human society better for everyone, and that artificial intelligence should be used to sustain the planet. This is the main principle that should support the other five ethical AI principles (transparency and explainability, privacy, safety and security, fairness and justice, and accountability).

Technology was invented by human society for human society. However, what would happen if AI systems destroyed our planet, exacerbated inequality and poverty, and eliminated all white-label jobs? This article will be devoted to the topic of artificial intelligence posing an existential threat to human society, and artificial intelligence may take over all people and their impact on the environment in terms of labor.

Existential threats

The question of whether AI poses an existential threat is a topic of constant debate and speculation among experts and researchers. In dealing with this issue, a variety of points of view must be carefully considered.

Some individuals and organizations have expressed concerns about the potential risks associated with developing and deploying advanced AI systems. These concerns range from AI systems surpassing human intelligence and becoming uncontrollable or acting in ways harmful to humans, to AI being used for malicious purposes, such as autonomous weapons or mass surveillance.

In fact, CEOs of tech companies, AI researchers, and academic professors have all signed the statement that mitigating the risk of AI extinction should be a global priority, along with other social-scale risks such as pandemics and nuclear war. The full list of signatories can be found here: https://www.safe.ai/statement-on-ai-risk#signatories

The statement comes amid growing concerns about the potential negative impact of AI. Notable advances in large language models such as ChatGPT and other chatbots have raised concerns about the widespread spread of misinformation and propaganda and the potential replacement of numerous white-collar jobs.

Advances in AI must continue to be monitored and studied, potential risks identified, and proactively addressed through technical, ethical and policy measures. This includes ongoing AI safety research, robust testing and validation of AI systems, and establishing frameworks to ensure responsible and beneficial AI deployment.

Here's what Munch debates about AI R&D as an existential threat.

labor

According to a report by Goldman Sachs, AI has the potential to replace about 300 million full-time jobs. With the continuous advancement of artificial intelligence technology, they are reshaping the pattern of human labor, and opportunities and challenges coexist. Automation is one of the main impacts of AI on human labor. AI systems excel at performing repetitive and routine tasks quickly, accurately, and efficiently, increasing business productivity and saving costs. However, this automation can also lead to job loss, especially for workers engaged in manual labor or predictable, rules-based activities. Goldman Sachs' report highlights the vulnerability of the administrative and legal professions to automation.

Links: https://www.weforum.org/agenda/2023/05/jobs-lost-created-ai-gpt/

Despite the replacement of jobs, AI also has the potential to create new roles and opportunities. As routine tasks become automated, employees can shift their attention to higher-level, creative, and complex tasks that require unique human skills such as critical thinking, problem-solving, and emotional intelligence. AI can complement human labor, augmenting human capabilities and allowing workers to focus on tasks that require humanity. The World Economic Forum identified several fastest-growing positions, including AI and Machine Learning Specialist, Sustainability Specialist, Business Intelligence Analyst, Information Security Analyst, Fintech Engineer, Data Analyst and Scientist, Robotics Engineer, Electrical Engineer, Agricultural Equipment Operator, and Digital Transformation Specialist.

In addition, AI technology can improve job performance and support workers to complete tasks. AI-driven tools and software help professionals with data analysis, decision-making, and problem-solving by providing valuable insights, predictions, and recommendations. This collaboration between humans and AI systems promotes efficiency, accuracy, and innovation in various fields. It is important to note that the impact of AI on human labor is not uniform across all sectors and occupations. Industries that involve complex social interactions (i.e., social workers), creativity (i.e., circus artists), adaptability, and emotional intelligence are less susceptible to automation. Similarly, roles that require physical flexibility, empathy (i.e., psychologists), and expertise may still be primarily human-driven.

Widespread adoption of AI requires a shift in workforce skills and capabilities. As automation takes over routine tasks, the demand for individuals who are well-versed in data analysis, machine learning, AI development, and algorithmic thinking increases. Upskilling and retraining programs are critical to equipping the workforce with the necessary capabilities to thrive in an AI-driven economy. Adapting to the changing employment landscape is critical for individuals to remain competitive and find new opportunities in the age of AI.

environment

AI is a rapidly evolving field with great potential for innovation and transformation in various fields. However, it is important to recognize that AI also has environmental impacts that need to be carefully considered. While AI can improve efficiency and sustainability in some areas, it may also lead to increased energy consumption and resource use.

In a new paper, researchers at the University of Massachusetts Amherst conducted a life cycle assessment of training several common large AI models. They found that the process could emit more than 626,000 pounds of carbon dioxide equivalent — almost five times that of the average U.S. car, including the manufacture of the car itself. In addition, AI applications that rely on massive data processing and storage consume a lot of power in data centers. These data centers require cooling systems to maintain optimal operating temperatures, resulting in additional energy consumption.

A major aspect of AI's impact on the environment is the computing power required to train and run AI models. Training complex deep learning algorithms requires significant computing resources, which often involve energy-intensive processes and powerful hardware infrastructure. This results in a large carbon footprint, especially in the case of large-scale AI deployments. Notably, ChatGPT alone consumes the equivalent of 175,000 people, and its third version of the training emits 550 carbon dioxide. In addition, companies like Google and Microsoft have not publicly disclosed their energy consumption to generate AI models.

However, it is worth mentioning that AI also offers opportunities for environmental sustainability. AI-driven optimization algorithms can enhance resource allocation and efficiency in various industries such as energy management, transportation, and agriculture. For example, AI-driven systems can optimize energy consumption in buildings, improve traffic flow to reduce congestion and emissions, and enable precision agriculture to minimize water and fertilizer use. The application of artificial intelligence in solving environmental challenges is shown in an interesting article linked here:

https://www.unep.org/news-and-stories/story/how-artificial-intelligence-helping-tackle-environmental-challenges