TY - JOUR

T1 - Quo Vadis? Стратегические ориентиры в развитии искусственного интеллекта и необходимость новой социальной аналитики

AU - Трегубова, Наталья Дамировна

AU - Резаев, Андрей Владимирович

PY - 2025

Y1 - 2025

N2 - The article explores the potential of new strategic guidelines to harness the power of artificial intelligence (AI) to better society’s everyday life and mitigate its risks. The paper’s opening section presents a detailed introduction and critical evaluation of two influential paradigms in the development of AI instruments. Both paradigms are deeply embedded in the traditions of Western European philosophical and economic thought. The first paradigm makes an orientation toward the concept of rational agents, designed with specific, predetermined goals that guide their behavior and decision-making. In contrast, the second paradigm focuses on emulating human capabilities through advanced computer technology, striving to mirror the complexity and nuance of human thought and action. In the second portion of the article, the authors examine Stuart Russell’s innovative approach to artificial intelligence technology. Renowned as a foremost expert in contemporary computer science, Russell advocates for developing “human-compatible AI”, a concept emphasizing the importance of aligning AI systems with human values and safety. His work explores the intricacies of creating intelligent machines that perform tasks effectively and operate harmoniously within societal frameworks. In what follows, the authors explore the opportunities and challenges associated with this approach. They look at the nuanced “red lines” suggested by scientists, which outline the ethical boundaries and considerations essential for responsible AI development. A thorough examination of both existing and prospective approaches to crafting AI algorithms uncovers the substantial obstacles that developers and managers encounter. These challenges are critical as they strive to safeguard humanity from the potential unintended consequences and harm that these advanced technologies could inflict. In conclusion, the article argues that the current approach, where society develops various versions of AI and only then considers regulation, is deeply flawed. The authors advocate for a shift in focus towards creating AI that is inherently compatible with humans and can be integrated into a specific society. They stress the urgent need for new social analytics to study the challenges of AI tool integration in the context of the evolving ‘human-algorithm’ and ‘algorithm-algorithm’ relationships.

AB - The article explores the potential of new strategic guidelines to harness the power of artificial intelligence (AI) to better society’s everyday life and mitigate its risks. The paper’s opening section presents a detailed introduction and critical evaluation of two influential paradigms in the development of AI instruments. Both paradigms are deeply embedded in the traditions of Western European philosophical and economic thought. The first paradigm makes an orientation toward the concept of rational agents, designed with specific, predetermined goals that guide their behavior and decision-making. In contrast, the second paradigm focuses on emulating human capabilities through advanced computer technology, striving to mirror the complexity and nuance of human thought and action. In the second portion of the article, the authors examine Stuart Russell’s innovative approach to artificial intelligence technology. Renowned as a foremost expert in contemporary computer science, Russell advocates for developing “human-compatible AI”, a concept emphasizing the importance of aligning AI systems with human values and safety. His work explores the intricacies of creating intelligent machines that perform tasks effectively and operate harmoniously within societal frameworks. In what follows, the authors explore the opportunities and challenges associated with this approach. They look at the nuanced “red lines” suggested by scientists, which outline the ethical boundaries and considerations essential for responsible AI development. A thorough examination of both existing and prospective approaches to crafting AI algorithms uncovers the substantial obstacles that developers and managers encounter. These challenges are critical as they strive to safeguard humanity from the potential unintended consequences and harm that these advanced technologies could inflict. In conclusion, the article argues that the current approach, where society develops various versions of AI and only then considers regulation, is deeply flawed. The authors advocate for a shift in focus towards creating AI that is inherently compatible with humans and can be integrated into a specific society. They stress the urgent need for new social analytics to study the challenges of AI tool integration in the context of the evolving ‘human-algorithm’ and ‘algorithm-algorithm’ relationships.

UR - https://www.mendeley.com/catalogue/6f52a691-843d-314f-a941-e592c1d627be/

U2 - 10.17223/1998863x/85/10

DO - 10.17223/1998863x/85/10

M3 - статья

SP - 112

EP - 122

JO - ВЕСТНИК ТОМСКОГО ГОСУДАРСТВЕННОГО УНИВЕРСИТЕТА. ФИЛОСОФИЯ. СОЦИОЛОГИЯ. ПОЛИТОЛОГИЯ

JF - ВЕСТНИК ТОМСКОГО ГОСУДАРСТВЕННОГО УНИВЕРСИТЕТА. ФИЛОСОФИЯ. СОЦИОЛОГИЯ. ПОЛИТОЛОГИЯ

SN - 1998-863X

IS - 85

ER -