#035 ML Григорий Бакунов (Бобук). Мораль, этика и философия искусственного интеллекта

#035 ML Григорий Бакунов (Бобук). Мораль, этика и философия искусственного интеллекта

#035 ML Grigory Bakunov (Bobuk). Morality, ethics and philosophy of artificial intelligence

by Machine Learning Podcast · View the channel on PSI

Humanities, Social Sciences & Thought AI & Society UBDigital and Information technology: general topicsUBJDigital and information technologies: social and ethical aspects
🎙 Machine Learning Podcast 👥 5K 📅 December 6, 2021 ⏱ 63 min 👁 2K 📄 expert opinion 🧭 2026-10-09
Available in: English (current) Français

Keywords

ethicsAIphilosophymoralitytrolley problem

Summary

In this episode of the Machine Learning Podcast, host Mikhail welcomes Grigory Bakunov (Bobuk), a prominent Russian IT expert, to discuss the ethics and philosophy of artificial intelligence. Bakunov recounts his early career in face recognition systems in the late 1990s and his subsequent journey into machine learning. He describes a current project funded by a European country aimed at building an AI system that resolves ethical dilemmas better than the average European. The system is trained on a dataset of about 100,000 crowdsourced answers from 14 European countries, with a goal to reach 500,000 answers. Bakunov discusses the trolley problem and argues that AI can learn from aggregated expert opinions rather than seeking absolute truths. He also addresses the need for explanations that are convincing, citing Pelevin’s novel as inspiration. The conversation then shifts to the future of AI: Bakunov believes the field is in a temporary impasse due to hardware limitations, and he looks to neuroscience, particularly Neuralink, for potential breakthroughs. He also criticizes the medical field for its rigid barriers to AI adoption. The episode ends with a discussion on how AI will impact creative industries like game development.

192 words

Critical Evaluation

Value of the Information & Strength of the Argument

The value of the information lies in the guest’s unique perspective as an engineer who has worked on AI for decades. His arguments are generally coherent, though they rely heavily on personal anecdotes and speculative reasoning. He offers a pragmatic view on AI ethics, suggesting that consensus of experts is more practical than philosophical absolutes. However, the lack of empirical data and the unverified nature of his project weaken the rigor. The discussion is thought-provoking but not deeply technical, with arguments that might oversimplify complex ethical challenges.

Scientific Rigor, Source Quality, Title Accuracy

The guest frequently references his own experiences and a few external works (e.g., Pelevin’s novel) but does not provide citations to academic sources. The description includes links to the podcast’s distribution channels and his Telegram channel, but no direct references to scientific studies are offered. The title is appropriate and matches the content, though the term ‘philosophy’ might be used loosely. Overall, the scientific rigor is low due to the conversational nature and lack of verifiable evidence.

179 words

Title / Content Match

The title accurately reflects the content: the conversation focuses on morality, ethics, and philosophy of AI, with examples and personal viewpoints from Grigory Bakunov.

Quality & Reliability

6/10

The discussion is based on the guest's personal experience and opinions rather than peer-reviewed research. While the guest is a recognized IT expert, his claims about AI ethics are anecdotal and lack systematic evidence. The ongoing project described is not yet publicly documented, so its viability remains unverified.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

External References

Contribution & Novelties

This podcast episode brings a practitioner’s perspective on AI ethics, emphasizing the construction of a crowd-sourced ethical decision system. The originality lies in the practical attempt to collect large-scale human responses to ethical dilemmas and train an AI on them, rather than purely theoretical discussions. The guest also speculates on future breakthroughs from neuroscience and brain-computer interfaces.

Pour aller plus loin :

95 words

Radar Profile

The scores indicate a moderate level of information quantity but lower technical depth and reliability, typical of an opinion-driven podcast. The profile suggests a blend of personal insights and general knowledge rather than rigorous scientific analysis.

Reliability 5/10