Since the evolution of computational and communication technologies, the world becomes a global village: society and environment are seen through digital perspectives of faster and wider industrial revolutions that become widely available and influence daily our lives. Generative AI dynamically trained and used on billions of blurred knowledge graphs’ multimodal elements becomes a growing technological presence with impact. New scientific and regulatory debates address, alongside the opportunities of generative AI presence in economy and society, the challenges that could, on one side, reinvent technologies and the humanity through the Singularity hypothesis, and one the other side, open its availability for unethical, improper, or illegal use of „black box” learning. Within dynamically challenging tsunami of big data, information and knowledge systems, the need of contemporary wisdom captured within AI could benefit from the well-known and visionary ethical yet traditional philosophical foundations. The need to adapt the technology based on well-known and accepted logics to unsafe, opaque decisions finds solid foundations in philosophical ethics. This paper critically reviews recent scientific articles on AI challenges, the Kantian perspective, and ethics foundations that argue the need of human in the loop for the current progresses of generative AI. The role of Kantianism in the ethical perception of AI and its accent on reason and reasoning for ethics, as well as other traditional ethical theories such as theory of virtue and utilitarianism, arguably request the presence of human values at the centre of AI and justify the argument of this paper that current challenges of transparency, explainability, fairness, responsibility of AI technologies can be addressed by consistent transitions from data-centred technologies to human-centred and scientifically quantifiable solutions for the next industrial revolution(s). The paper includes case studies of ethical and common-sense challenges for the use of „black box” models in healthcare, paediatric robotics, also refers to other potential applications such autonomous decision making or creativity. The paper ends with arguments and open questions on opportunities and challenges for AI and philosophy scientists with a focus on ethical aspects of AI technologies.
Read full abstract