AI如何处理极端问题

However,AI can increase the riskof unfair bias or discrimination across a range of indicators or characteristics.Thiscould undermine public trust in AI.Product safety laws ensure that goods manufactured and placed on the market inthe UK are safe.Product-specific legislation(such as for electrical and electronicequipment,56medicaldevices,57and toys58)may apply to some products thatinclude integrated AI.However,safety risks specific to AI technologies should bemonitored closely.As the capability and adoption of AI increases,it may pose newand substantial risks that are unaddressed by existing rules.Global Innovation Index 2022,GII 2022; Global Indicators of Regulatory Governance,World Bank,2023.Demand for AI skills in jobs,OECD Science,Technology and Industry Working Papers,2021.The protected characteristics are age,disability,gender reassignment,marriage and civil partnership,race,religion orbelief,sex,and sexual orientation.Article 5(1)(a)Principles relating to processing of personal data,HM Government,2016.Electrical Equipment(Safety)Regulations,HM Government,2016.Medical Devices Regulation,HM Government,2002.Toys(Safety)Regulations,HM Government,2011.A pro-innovation approach to AI regulationConsumer rightslaw59may protect consumers where they have entered into asales contract for AI-based products and services.Certain contract terms(forexample,that goods are of satisfactory quality,fit for a particular purpose,and asdescribed)are relevant to consumer contracts.Similarly,businesses are prohibitedfrom including certain terms in consumer contracts.Tort law provides acomplementary regime that may provide redress where a civil wrong has causedharm.It is not yet clear whether consumer rights law will provide the right level ofprotection in the context of products that include integrated AI or services based on

在上面这个非常简单的例子中，我们很容易能够显式地生成整个多路图。但在大多数实际示例中，该图会太大。因此，挑战通常是在不追踪整个可能性图的情况下弄清楚要采取什么行动。一种常见的方法是尝试找到一种方法来为不同的可能状态或结果分配分数，并仅追求分数最高的路径。在自动定理证明中，“从初始命题向下”和“从最终定理向上”工作也很常见，试图找出路径在中间的交汇处。还有另一个重要的想法：如果建立了“引理”，即存在从X到Y的路径，则可以将X→ Y添加为规则集合中的新规则。So how might AI help?As a first approach,we could consider taking something like our string multiway system above,and training what amounts to a language-model AI to generate sequences of tokens that represent paths(or what in a mathematical setting would be proofs).The idea is to feed the AI a collection of valid sequences,and then to present it with the beginning and end of a new sequence,and ask it to fill in the middle.那么人工智能可以提供什么帮助呢？作为第一种方法，我们可以考虑采用类似于上面的字符串多路系统的东西，并训练相当于语言模型人工智能的东西来生成代表路径的标记序列（或者在数学设置中将是证明）。这个想法是向人工智能提供一组有效的序列，然后向它呈现一个新序列的开头和结尾，并要求它填充中间部分。

address the challenges created by these characteristics,we future-proof our framework againstunanticipated new technologies that are autonomous and adaptive.Because we are notcreating blanket new rules for specific technologies or applications of AI,like facial recognitionor LLMs,we do not need to use rigid legal definitions.Our use of these defining characteristicswas widely supported in responses to our policypaper,81as rigid definitions can quickly becomeoutdated and restrictive with the rapid evolution ofAI.82We will,however,retain the ability toOne of the biggest problems in regulating AI is agreeing on a definition,Carnegie Endowment for International Peace,2022.Establishing a pro-innovation approach to regulating AI,Office for Artificial Intelligence,2022.As stated in government guidance on using AI in the public sector,we consider machine learning to be a subset of AI.Whilemachine learning is the most widely-used form of AI and will be captured within our framework,our adaptive and autonomousA pro-innovation approach to AI regulationadapt our approach to defining AI if necessary,alongside the ongoing monitoring and iteration ofthe wider regulatory framework.