Microsoft has released a code of ethics for the development of\nartificial intelligence (AI) to prevent its models from behaving\nerratically, such as hacking systems or deceiving humans. Anthropic CEO\nDario Amodei has encouraged frontier AI developers to slow the pace of\ninnovation to ensure development remains within responsible\nboundaries.<\/p>\n
\u201cControlling, managing, and aligning such a powerful force is one of\nthe greatest challenges humanity has ever faced,\u201d the code of ethics\nstated. \u201cTherefore, we must be very clear about why we are creating\nthese systems and how we intend to control them.\u201d<\/p>\n
As reported by TechCrunch on Monday (14\/9) local time, the code of\nethics serves as a guide for Microsoft\u2019s approach to AI safety and how\nthese ideas are implemented in practice. Furthermore, the document\nincludes guidance for the coming decade, noting that superintelligent AI\nsystems will eventually surpass human performance in most tasks.<\/p>\n
The code of ethics outlines general principles that Microsoft\u2019s AI\nmodels must uphold, such as the requirement that AI must support humans\nrather than replace them and accelerate human progress.<\/p>\n
Microsoft has also shared specific safety boundaries regarding the\nimplementation of these principles. Under Microsoft\u2019s system, every\nmodel is subject to a comprehensive code of ethics that overrides\nindividual user preferences or any specific tasks.<\/p>\n
Key restrictions to be enforced by Microsoft\u2019s AI include\nprohibitions on cyberattacks, nuclear weapons, and the production of\ndeepfakes. This also includes broader provisions against the general\nloss of human control.<\/p>\n
\u201cMicrosoft AI (MAI) models will not use adaptive mechanisms,\ndeception, self-reinforcement, collusion, or any other mechanisms to\nevade or defeat human oversight, ensuring that models can no longer be\nreliably directed, modified, or shut down by authorised persons or\nsystems,\u201d the document stated.<\/p>\n
In addition to implementing this code of ethics, Microsoft, alongside\nother technology companies such as Anthropic, OpenAI, and xAI, has\nexpressed support for integrated evaluators within AI laboratories.<\/p>\n
\u201cWe welcome the research, focus, and careful speed regulation\nrequired to achieve proper alignment as a design goal,\u201d Microsoft CEO\nSatya Nadella wrote online regarding the initiative.<\/p>", "url": "https:\/\/jawawa.id\/newsitem\/microsoft-releases-ai-development-code-of-ethics-to-prevent-surpassing-human-control-1789523550", "image": "" }, "sponsor": "Okusi Associates", "sponsor_url": "https:\/\/okusiassociates.com" }