AI and society: Implications for global equality and quality of life

Cognitive capabilities have long differentiated humans from other species and machines. Thinking, reasoning, learning, solving complex problems and making critical decisions set humans apart. Not anymore. “Attention is all you need,” a research paper published in 2017 by scientists at Google, detailed the advent of generative AI, which replicates many humanlike cognitive capabilities and, in some cases, outperforms our ability to conduct complex, cognitive-type tasks. AI has no inherent sense of morality; it simply extracts and extrapolates on correlations or patterns in data. Yet, its ability to mimic cognitive capabilities places it in a privileged position to interpret and act based on our morally informed — and sometimes biased — instructions.

Creativity, on the other hand, involves using foundational cognitive skills (e.g., perception, memory, attention, visual processing) to come up with novel ways of thinking or doing things, often making connections between seemingly unrelated concepts. Tests run by researchers (e.g., the Torrance tests of creative thinking, alternative uses test and remote associates test) suggest that generative AI can materially outperform humans in creative tasks and ideation, even though the generated content is based on existing data. 

Using generative AI as a creative co-pilot can facilitate inspiration, help in brainstorming and connect ideas to spark the creative process and overcome blank canvas syndrome. 

S&P Global tested generative AI's ability to creatively connect seemingly unrelated topics with an experiment. We asked ChatGPT to generate a concept for a home-delivery business based on three unrelated words: nuts, Greece and hair. The result was a Greek, nut-infused hair care product line that would be sold online and delivered to customers’ homes. When we further prompted the AI to adopt an “entrepreneur persona,” it produced a list of unique selling points and a target market. Regardless of the idea’s marketability or financial sense, the experiment offers evidence of AI's ability to augment human cognitive capabilities and creativity. 

Generative AI may also help democratize innovation (as defined in Eric von Hippel’s 2005 book, Democratizing Innovation) as the technology can be applied to the whole creative-production process from ideation to evaluation, idea selection, risk and impact analysis, project feasibility, project execution, and regulatory compliance. When used in this way, AI can facilitate human creativity, not least by augmenting our ability to manage the necessary, noncreative elements that accompany creative processes.

Yet AI's creative capability could have detrimental impacts if it results in the replacement or muting of human creativity. Much like the calculator prompted the widespread demise of mental calculations, generative AI could disincentivize curiosity and divergent thinking, leading to a decline in humanity's pursuit of creative endeavors. If AI broadly replaces humans in industries such as entertainment, professional services and art (while also discouraging students from pursuing creative studies), it could diminish and ultimately replace the pool of human creative talent. That would have negative implications for the diversity and richness of human creativity, exploration and curiosity. 

A growing reliance on generative AI for ideation will also raise intellectual property issues, including attribution and copyright. For example, current regulations do not afford patent or copyright protection to machine-generated content; these protections are limited to human creators.

Those risks demand that generative AI's role in cognitive and creative endeavors is subjected to rules that preserve humanity’s intrinsic motivation and curiosity. The increasing use of AI presents a novel challenge for regulators and educational institutions (here broadly defined to include schools, universities, professional education bodies and corporations) to incentivize and guide a critical mindset and creative thinking skillset across society. 

Generative AI has the potential to improve people's lives by streamlining the mundane and quotidian (e.g., reducing customer service waiting times and taking over time-consuming administrative tasks) as well as personalizing a broad range of services (e.g., shopping, travel and leisure). But AI’s most significant societal impact may be in the ways that it materially improves access to and the quality of essential goods and services, including healthcare, education and food.

Healthcare: AI stands to improve services and outcomes at every stage of the healthcare life cycle, including research and innovation, prevention, advanced diagnostics and imaging, treatments, patient experience, recovery, and administration efficiency (see “AI in Healthcare: A Path to Long-term Immunity?” June 2024). AI-enhanced testing and scenario analyses could accelerate new drug discoveries and reduce time to market. Advanced diagnostics using AI can be more effective than the human eye in identifying hidden patterns, leading to earlier and more accurate diagnoses. Surgeries may become more effective and less invasive with AI-assisted advances in robotics and procedures. Patients may experience more personalized care, improved monitoring, reduced errors, quicker recovery, better access to preventive treatments and more streamlined interactions with health insurance providers. Hospital systems may better optimize deployment of medical staff while reducing administrative costs. 

AI could also make advanced medical care accessible to previously underserved, low-income and rural communities. AI-assisted telemedicine, for instance, can bring more advanced and higher-quality medical care into geographically isolated medical facilities and households. Virtual consultations with specialists and remote diagnostics free those in need of care from the limitations of local resources and facilities. New and faster drug discovery and reductions in health administration costs can make quality care affordable to a wider cross-section of society. 

Education: Generative AI may be most immediately (and negatively) thought of for its ability to write students’ papers. But AI also has vast potential to advance and enhance learning. Lower-income and struggling students are already gaining access to AI-powered tutoring and customized learning programs that provide the type of support once reserved for students with significant financial resources. Teachers should benefit from help developing tailored learning paths and assessments based on their classes' needs, while AI-assisted lesson planning will free up educators’ time to focus on their students, reducing unpaid overtime that can lead to burnout and increased staff turnover. AI can also assist in the development of advanced training materials, such as simulations, to better prepare individuals for careers requiring specialized skills. 

Food: Genetic modification has been used for years to improve the taste, appearance and other qualities of agricultural products, including fruits, vegetables and grains. Synthetic biology (synbio) powered by AI will expand the creation and modification of organisms, including those central to the global food supply (see “Artificial Intelligence Powering Synthetic Biology: The Fundamentals,” June 2024). Synbio techniques can be used to directly alter plant and animal DNA, providing greater flexibility in creating new traits. Food technology companies are increasingly focused on improving food and water safety, nutrition levels, and resilience to complex climate conditions, aligning them with some of the UN’s sustainable development goals. AI’s growing role in this field should improve scalability and could ultimately help reduce world hunger, food costs and production time. 

As with all aspects of AI integration, stakeholders must also carefully consider risks that could lead to negative outcomes for segments of society. In healthcare, for instance, there are legitimate concerns over data protection, inherent algorithmic biases, medical decision-making transparency and accountability. In education, there are concerns about the reliability of information accessed through AI, that overreliance on technology could inhibit independent thought and that AI could reinforce existing biases. There are also wider risks to using AI-enabled shortcuts to obtain knowledge. Indeed, it is often the journey of learning — asking questions and seeking answers — that assists in the absorption and contextualization of knowledge and builds learning skillsets. For the food industry, the introduction of new edible organisms might lead to unintended novel allergies or intolerances, potentially resulting in collateral effects on animal and human health that are difficult to foresee.

Although AI integration could vastly expand and democratize the provision of important quality-of-life services to underserved communities, particularly in the long term, we must also consider distributional effects. For example, will the use of AI widen the existing gaps between those with and without access?

AI could transform some of the world's largest and most labor-intensive industries. Yet its adoption and resulting impact, both across and within societies, will likely be uneven. The differences that emerge could affect global growth and the nature of employment (see "Investment And Talent Are The Keys To Unlocking AI’s Potential," July 2024). Depending on how it is applied, AI could drive greater societal equity or widen inequality. Factors such as infrastructure, tech-savviness, skilled labor markets, and countries' differing resources and willingness to harness AI will determine the technology’s ultimate impact on equality between and within countries.  

We expect advancements in generative AI will disrupt labor markets by complementing some jobs, creating new ones and replacing others. Broad implementation of generative AI could affect income equality (through wages and productivity) and wealth equality (where stakeholders benefit from companies creating AI-related value-adds or where populations benefit from better access to AI-related education). GDP per capita in emerging markets is projected to be about a third of that in developed markets by 2030 (not too different from today’s); therefore, to accelerate convergence, productivity growth must improve in emerging markets. However, increasing automation and AI adoption are arguably the biggest threat to emerging markets in the coming decade, since the use of relatively cheap labor will likely be insufficient to compete, and there will be an increasing need to invest in technological skills and mechanized manufacturing to ensure continued supply chain-led growth (see “Look Forward Emerging Markets: A Decisive Decade,” October 2024).

We think that in the short to medium term, broad-based application of generative AI could lead to higher inequality within and between economies. This is particularly likely for parts of society in developed economies with relatively deep capital markets, large private sectors, supportive governments and qualified human resources. We expect earlier adoption by these societies will initially increase inequality between developed and developing economies. 

Approximately 2.6 billion people (about one-third of the world’s population) do not have internet access. This will act as an anchor on developing economies because most AI models today rely on internet access to function. The digital divide resulting from this inability to access and deploy AI technology will likely increase inequality between emerging and developed economies, at least in the near term. The likely effects of AI on labor markets and inequality are the subject of much debate and reflect diverging views. First and foremost, we think generative AI will be a productivity-enhancing tool that will augment human knowledge and abilities to perform complex tasks. In the medium term, AI models that can run offline — which are currently being developed and can leverage computing at the edge (e.g., on smartphones or laptops) as opposed to in the cloud — could increase equality across economies and decrease the gap between developed and developing economies. That said, the complex interplay of various global megatrends and national characteristics (e.g., primary industry sectors, companies, labor markets) means there will be continued uncertainty regarding AI's longer-term effects on inequality between and within societies. 

Each of the discussions on cognitive capacity, quality of life and economic equality touched on concerns regarding algorithmic biases, uneven cognitive and social impacts, and other potentially detrimental effects of AI’s integration into society. Despite these shortcomings and risks, we view the proliferation of AI technologies as inevitable. The question thus becomes the following: How can society effectively manage and mitigate these risks and maximize AI’s net beneficial potential? That responsibility lies largely in the hands of key institutions, namely the private sector, government regulators, and nongovernmental and transnational organizations. Moreover, it will depend on the governance structures they erect and enforce (see “The AI Governance Challenge,” Nov. 29, 2023). As Nobel Prize winner Joseph Stiglitz said, history demonstrates that “unfettered capitalism, unfettered innovation, does not lead to the general well-being of our society.”

Institutions will be key determinants of AI's net benefit, but the equation will also depend on changing social behaviors, societal norms and cultural acceptance of shifting dynamics as AI adoption proliferates. In that interplay, one of the tasks of public, private and nongovernmental institutions will be to catalyze AI innovation and regulate its application in ways that best serve equity and inclusion and minimize associated risks.