In a recent publication in the journal Futures, researchers explored the emergence of future scenarios in the age of artificial intelligence (AI) from the perspective of AI and robotics practitioners.
Study: Navigating AI's Uncertain Future: Practitioners' Perspectives. Image credit: Monster Ztudio/Shutterstock
Background
AI, a technology of profound transformation, autonomously learns, interacts with humans in real-time, and pervades work, the economy, and daily life. The recent introduction of AI-driven applications like ChatGPT, Midjourney, and DALL-E has further fueled the current AI frenzy. These tools enable interactions with AI for answering questions, generating images, and composing essays, producing credible yet potentially deceptive content. AI can also generate unexpected and inaccurate information. The integrity of knowledge, creativity, trust, and transparency, collectively known as the "knowledge economy," is at stake.
Predictions suggest generative pre-trained transformers (GPTs) could impact a significant portion of job tasks. Technology leaders have publicly voiced apprehensions in response to these concerns and the rapid global development of AI. In this context, researchers conducted an empirical study on future perspectives on AI.
Imaginaries of the AI era
AI sparks the imagination, giving rise to diverse utopian and dystopian visions of its era. Dystopian AI imaginaries depict human-machine conflict, job displacement, and weaponization, while utopian visions envision AI-human coexistence and empowerment. These imaginaries represent collective and relatively stable future concepts.
To enrich the understanding of AI futures, a nuanced vocabulary is needed. This involves recognizing the continuum between utopian and dystopian elements and considering multiple interpretations and non-linear trajectories. Individual perceptions of desirable and undesirable futures influence socio-technical future imaginaries shaped by collective visions of the AI era. These imaginaries inherently involve tensions, representing arguments for and against specific future scenarios shaped by decisive questions.
Spectrum of AI future imaginaries
The empirical data consists of 35 interviews conducted with AI and robotics practitioners. These interviews encompassed 26 individuals working in various industries and organizations engaged in AI development and application, along with nine participants from university settings specializing in robotics. The motivations and goals of practitioners in academia and industry differed substantially.
In universities, researchers primarily drove theoretical exploration and innovation, actively engaging in broader societal discussions about AI's future. In contrast, industry developers actively pursued commercial applications and actively disseminated the AI technologies they developed. Despite these divergent motivations, both academic and industrial practitioners communicated their visions of AI futures to a broader society.
The study adopts a qualitative research design to gain insight into the content and emergence of future imaginaries in the AI era from the perspectives of AI and robotics practitioners. Qualitative interviews were chosen over larger-scale surveys to enable a more in-depth exploration of participants' thoughts and to provide a richer understanding of their worldviews. While the sample size of 35 AI practitioners is relatively small, it allows for the examination of less-explored "peripheral voices" in the field.
The analysis revealed three distinct continuums of best and worst future imaginaries, which practitioners described in relation to each other. These continuums encapsulate various aspects of AI futures, including vague utopian-dystopic scenarios, domain-specific descriptions, and economic and power-related considerations.
The first continuum involves a broad and vague vision of human-AI coexistence, encompassing both optimistic and dystopian elements. On this continuum, AI is envisioned as a helpful assistant that takes over repetitive work, granting people more time for meaningful, creative, and social pursuits. However, the dystopian extreme portrays AI as an evil force that enslaves humanity, leading to a virtual existence and the loss of human agency.
The second continuum focuses on domain-specific imaginaries, where the best future involves human-AI co-created innovations supporting sustainability, knowledge advancement, and societal improvement. Practitioners on this continuum emphasize the potential of AI to benefit society and envision a democratic, equitable world. However, the worst scenario in this continuum depicts increased inequality, AI-driven manipulation, and resource consumption, with people using AI for exploitative purposes.
The third continuum delves into economic and power-related aspects of AI futures. In the best scenario, practitioners envision a sustainable economy with dispersed power, regulated data usage, and AI contributing to ethical and sustainable development. However, the worst scenario entails corporate dominance over AI development, concentration of power, and increased societal inequality.
The research uncovers pivotal questions by examining the tension between the best and worst future scenarios. First, the issue of agency arises, questioning whether humans, AI, or both hold control. The potential for dystopian futures emerges if AI acts independently. Then, the concept of co-agency between humans and technology is explored, offering a positive outlook. However, it necessitates a delicate balance to prevent complete dependence on technology.
The second critical question concerns societal equality, linked to AI development's impact on society and the practitioner community. A diverse practitioner group is seen as more likely to foster democracy. Finally, the concentration of power in the global AI community and economy is questioned, with concerns about data monopolies and the need for government regulation.
Conclusion
In summary, researchers delved into the future of AI through 35 interviews with practitioners in Finland and Singapore. It reveals diverse future imaginaries, spanning utopic to dystopic continuums. These imaginaries prompt crucial inquiries about control, societal implications, and power distribution. Additionally, the study underscores the need for a redefined perspective on AI's role in shaping our future.
Meta Llama 3 Enhances Text-to-SQL Accuracy With Advanced Prompt Engineering and AWS Integration