The AI-Control-Problem must now be the Top Priority
The AI-Control-Problem must now be the Top Priority - Murat Durmus

The AI-Control-Problem must now be the Top Priority

Murat Durmus Murat Durmus

Murat Durmus

CEO & Founder @ AISOMA AG | Thought-Provoking Thoughts on AI | Member of the Advisory Board AI Frankfurt | Author of the book "MINDFUL AI" | AI | AI-Strategy | AI-Ethics | XAI | Philosophy

发布日期: 2023年1月31日
+ 关注

The AI control problem refers to ensuring that artificial intelligence (AI) systems act in ways that are consistent with human values, goals, and interests. This complex problem involves understanding and mitigating the potential risks associated with AI while ensuring that these systems are trustworthy, safe, and ethical; To address the AI control problem, researchers and practitioners in the field of AI ethics and governance are working to develop the best practices, ethical principles, and technical solutions that can help ensure that AI systems behave in ways that are socially responsible and consistent with human values; This includes areas such as transparency, accountability, fairness, and explainability in AI systems and the development of AI systems that can learn from human feedback and preferences.

The following points should be considered:

  1. Safety:?The AI system should not cause harm to humans or the environment.
  2. Algorithm transparency:?Understanding how the AI system makes decisions should be possible.
  3. Bias:?The AI system should be trained on unbiased data to prevent discrimination and prejudice.
  4. Privacy:?The AI system should not compromise the privacy of individuals or organizations.
  5. Ethics:?The AI system should align with ethical and moral principles.
  6. Responsibility:?There should be clear accountability for the decisions made by the AI system.
  7. Sustainability:?The AI system should be designed to have a minimal environmental impact and to be energy efficient.
  8. Security:?The AI system should be protected from malicious attacks and breaches.
  9. Human-AI interaction:?The AI system should be designed for effective and safe human interaction.
  10. Interoperability:?The AI system should be compatible with other systems and technologies.


Murat

More thought-provoking thoughts:

Mindful AI: Reflections on Artificial Intelligence


领英推荐

Thought-Provoking Quotes & Reflections on Artificial Intelligence


New Book Release: Beyond the Algorithm: An Attempt to Honor the Human Mind in the Age of Artificial Intelligence (Wittgenstein Reloaded)


Deutsche Ausgabe:


Thoughts on AI (weekly) Thoughts on AI (weekly)

Thoughts on AI (weekly)

14,999 位关注者

订阅
Nikolaj Van Omme

Developing a better AI that optimise complex industrial problems by 20-40% in production!

2 年

It probably can not be solved within ML... but with OR this can be done...

赞
回复
1 次回应
POOJA JAIN

Storyteller | Linkedin Top Voice 2024 | Senior Data Engineer@ Globant | Linkedin Learning Instructor | 2xGCP & AWS Certified | LICAP'2022

2 年

Informative and important pick on #AI

赞
回复
2 次回应
Bill McLellan

HCLS Industry Principal | Senior Solutions Architect

2 年

Murat Durmus I've been wondering, could there be any usefulness to safe AI of combining linear/additive models where errors rebound/revert to mean vs exponential/multiplicative models where errors compound? Should AI that doesn't contain a human in the loop be regulated to use composite models containing both batch linear (safer, more explainable) and online exponential (more accurate, less explainable) models like reinforced learning so alerts go blaring when predictions, recommendations, or actions deviate significantly from the linear model over some period of time? Maybe we could catch the paperclip factory from turning the world into paperclips.

赞
回复
查看更多评论

要查看或添加评论，请登录

Murat Durmus的更多文章

  • On Writing
    2025年3月23日

    On Writing

    Writing is increasingly becoming a ritual, a performance rather than a personal expression. It's a kind of…

    3 条评论
  • The Gentle Cage of Certainty
    2025年3月16日

    The Gentle Cage of Certainty

    Refusing to ask questions means shutting ourselves away. Not into an iron cage, but into a gentle, familiar silence…

    3 条评论
  • Me Against an Armada of AI Agents
    2025年3月9日

    Me Against an Armada of AI Agents

    A single, defiant human..

    2 条评论
  • The Silent Surrender
    2025年3月2日

    The Silent Surrender

    The future of AI ethics will not be decided by committees drafting guidelines. It will be decided by the silent…

    13 条评论
  • The Slow Theft of Self
    2025年2月25日

    The Slow Theft of Self

    I feel exposed, like my mind is shedding layers I never agreed to lose. I am naked, not in body but in thought.

  • The Abyss of Knowing or When Expertise Feels Like Guesswork
    2025年2月23日

    The Abyss of Knowing or When Expertise Feels Like Guesswork

    Many know that AI is not just a technological development but an epistemological crisis. It forces us to question not…

    2 条评论
  • The Difference Between AI Safety, AI Ethics, and Responsible AI
    2025年2月16日

    The Difference Between AI Safety, AI Ethics, and Responsible AI

    Some worry about existential threats, others about fairness, and others want to avoid bad PR. AI security, AI ethics…

    5 条评论
  • From AI-Ethics to Algorithmic Conscience
    2025年2月9日

    From AI-Ethics to Algorithmic Conscience

    An algorithm can calculate probabilities, but a conscience weighs consequences. If we insist on calling it 'AI ethics,'…

    8 条评论
  • From AI Agents to AI Monads
    2025年2月3日

    From AI Agents to AI Monads

    The path from AI agents to AI monads reflects the philosophical development from Cartesian dualism to Leibnizian…

  • The Courage to be Uncertain
    2025年2月2日

    The Courage to be Uncertain

    The greatest enemy of certainty is courage, the courage to admit, "I don't know." This may sound like an excuse in a…

    6 条评论

社区洞察

其他会员也浏览了