Technical Writer, Frontier AI Best Practice

London办公方式未标注

全职Other

薪资未披露

签证支持雇主未明示 · 无官方资质名单可查

🛡 岗位真实性核验

Ghost Job 检测

🛡 申请链接核验有效

信号由后台自动核验生成，仅供参考；判断权在你。设计案：GHOST-JOB-P1

🤖 AI 中文导读 · 由岗位 JD 原文生成

🎯

必备技能

AIRisk Assessment内容运营安全

免费注册后解锁"你的匹配分 vs 高分画像"与个人技能缺口。

💼

职位描述 · 原文

<div class="content-intro"><h2>About the AI Security Institute</h2>
The AI Security Institute is the world's largest and best-funded team dedicated to understanding advanced AI risks and translating that knowledge into action. We’re in the heart of the UK government with direct lines to No. 10 (the Prime Minister's office), and we work with frontier developers and governments globally.
We’re here because governments are critical for advanced AI going well, and UK AISI is uniquely positioned to mobilise them. With our resources, unique agility and international influence, this is the best place to shape both AI development and government action.</div><div class="vac_display_field">
<div class="vac_display_field_value">
<h2 class="vac_display_field">The deadline for applying to this role is Sunday 29th November 2026, end of day, anywhere on Earth.</h2>
<h2>About the team</h2>
AISI's Technical Standards and Best Practice team identifies what best practice is needed to manage the most serious risks from frontier AI development and leverages AISI’s position to get it adopted. 
AI capabilities are advancing faster than the science and practices used to manage their risks. Risks now emerge earlier, during training, internal deployment and evaluation, not just at public release, and evaluation methods are struggling to keep pace. Yet there is no consensus on the set of practices needed to safely use and develop the most capable AI systems, and high-level principles are not enough to manage these risks. What's needed is concrete technical guidance that developers agree is necessary. 
AISI is uniquely placed to lead on parts of this industry-wide effort to establish best practice. We house researchers with state-of-the-art technical insight on keeping AI secure during internal use, hold existing relationships with industry and other frontier AI organisations, and offer a neutral third-party home to establish best practice collaboratively. This role directly delivers on part of the Prime Minister’s September 2026 commitment that the UK will work towards principles and standards that ensure the most capable AI systems are developed securely and with appropriate transparency. It builds directly on our October 2026 publication <a href="https://www.aisi.gov.uk/blog/towards-technical-best-practice-for-frontier-ai-development">Towards technical best practice for frontier AI development</a>, setting out our view of the most urgent priorities. 
This is a new, growing team with high ambitions. It is a cross AISI initiative that draws on and collaborates with researchers and engineers from across AISI's research teams, as well as our Communications, International and Strategy teams. We're now building a dedicated core team and expect it to grow as the work does. 
<h3>Why this team matters: </h3>
AISI is uniquely placed to shape this guidance, and this team is how we'll do it: 
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="61" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="1" data-aria-level="1">Technical expertise: AISI has roughly 10 technical and research teams whose expertise can be drawn on; one of the largest groups working on AI risk outside the frontier AI companies. AISI has evaluated more than 55 frontier models before release and published research on topics such as safety cases, monitorability and control. </li>
</ul>
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="61" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="2" data-aria-level="1">A research institute within government: we work directly with leading AI developers, <a href="https://www.nist.gov/news-events/news/2026/07/uk-aisi-caisi-preliminary-assessment-kimi-k3s-cyber-capabilities">other governments</a>, and third-party evaluators, and can bring them together around practices no single company would set on its own. UK AISI also coordinates the International Network for Advanced AI Measurement, Evaluation and Science – a group of 10 AI security institutes and equivalents – which has already published its own and socialised others’ best practices.  </li>
</ul>
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="61" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="3" data-aria-level="1">A track record of shaping industry practice: AISI has already helped establish voluntary practices such as pre-release testing, and built Inspect, now the standard framework for leading third-party evaluators. </li>
</ul>
We're not seeking to mandate requirements. We want to contribute our expertise to an industry-wide effort to develop technical best practice that could later inform formal standards. 
<h3>What we're working on:</h3>
Examples of our first priority areas include: 
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="61" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="4" data-aria-level="1">Overseeing AI systems: what monitoring and controls developers need to make sure AI systems behave as intended, and how to preserve that oversight as systems become more capable. </li>
</ul>
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="61" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="5" data-aria-level="1">Sharing information on the speed of AI development: which metrics best track progress, including AI's use in its own development, and when they should be shared with governments and the public. </li>
</ul>
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="61" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="6" data-aria-level="1">Assessing and documenting risk: what risk assessments and safety cases should cover, and what standard of evidence they should meet.</li>
</ul>
 
<div class="vac_display_field">
<h2>About the role</h2>
<div class="vac_display_field_value">
Primary objective: turn AISI's technical expertise into authoritative best-practice documents that frontier AI developers find valuable and use, and into clear explanations of why those practices matter. 
You'll work closely with AISI researchers and the standards team on two kinds of writing. 
<h3>Example responsibilities:</h3>
<ul>
<li>
Best-practice documents (about 50%) 
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="4" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="1" data-aria-level="1">Work with researchers to decide what form each piece of best practice should take, and how it should be structured. </li>
</ul>
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="4" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="2" data-aria-level="1">Own delivery of best-practice documents, co-drafting by managing researchers’ technical input and editing to ensure each practice is coherent, specific, testable and technically accurate. </li>
</ul>
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="4" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="3" data-aria-level="1">Set out the technical rationale for each practice, drawing on research findings, threat models and evaluation results, so the case for it is evidence-based and credible to a diverse set of expert readers.  </li>
</ul>
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="4" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="4" data-aria-level="1">Develop a consistent style, templates and terminology for AISI best practice documents, and run review rounds with internal and external experts.</li>
</ul>
</li>
<li>
 Explainers (about 50%) 
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="4" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="5" data-aria-level="1">Write public explainers and technical blog posts setting out the reasoning behind AISI's best practice: why it matters, and what it means for developers, policymakers and the public. For example, the series that follow our agenda, <a href="https://www.aisi.gov.uk/blog/towards-technical-best-practice-for-frontier-ai-development">Towards technical best practice for frontier AI development.</a> </li>
</ul>
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="4" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="6" data-aria-level="1">Manage multi-stakeholder drafting processes, engaging with feedback from industry, wider government and expert third parties. </li>
<li data-leveltext="" data-font="Symbol" data-listid="4" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="7" data-aria-level="1">Adapt material for different audiences, from technical researchers to non-technical policy readers, working with AISI's communications team.</li>
</ul>
</li>
</ul>
<h3>Reporting and growth:</h3>
You'll report to the Head of Technical Standards and Best Practice. This role by default has no line management responsibilities thought responsibility will scale as the team grows. You'll learn from AISI's researchers across a wide range of frontier AI safety topics, and help shape a new team as it grows.
 
<h2>What we are looking for</h2>
We're flexible on the exact profile and expect successful candidates will meet many, but not necessarily all, of the criteria below. This role is most likely to fall at Level 4, but we are open to exceptional candidates at other levels. 
<h3>Essential:</h3>
We're looking for someone who can: 
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="2" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="1" data-aria-level="1">Understand nuanced technical AI safety work well enough to explain its claims and limitations accurately, </li>
</ul>
<ul>
<li data-leveltext="" data-font="Symbol" data-listid="2" data-list-defn-props="{"335552541":1,"335559685":720,"335559991":360,"469769226":"Symbol","469769242":[8226],"469777803":"left","469777804":"","469777815":"hybridMultilevel"}" data-aria-posinset="2" data-aria-level="1">Write precise, structured documents that set out what organisations should do and why,<span data-ccp-props="{"134245418":true,"335559738":200,"335559739&

🈶

中文翻译 · AI 生成，以原文为准

<div class="content-intro"><h2>关于人工智能安全研究所</h2>
人工智能安全研究所是全球最大、资金最充裕的团队，致力于理解高级人工智能风险并将这些知识转化为行动。我们位于英国政府中心，与首相办公室（第10号）直接联系，并与全球前沿开发者和政府合作。
我们存在的原因是政府对于高级人工智能的顺利发展至关重要，英国AISI独特地处于动员他们的位置。凭借我们的资源、独特的敏捷性和国际影响力，这是塑造人工智能发展和政府行动的最佳场所。
</div><div class="vac_display_field">
<div class="vac_display_field_value">
<h2 class="vac_display_field">此职位申请截止日期为2026年11月29日星期日，地球上的任何地方。</h2>
<h2>关于团队</h2>
AISI的技术标准和最佳实践团队确定需要哪些最佳实践来管理前沿人工智能发展的最严重风险，并利用AISI的位置推动其采用。
人工智能能力的发展速度超过了管理和这些风险所使用的科学和实践。现在风险在培训、内部部署和评估过程中出现，而不仅仅是公共发布时，评估方法也难以跟上。然而，对于确保安全使用和开发最强大的人工智能系统的实践集合尚未达成共识，高级原则也不足以管理这些风险。需要的是开发人员一致认为必要的具体技术指导。
AISI独特地处于领导这一行业共同努力，以建立最佳实践的部分位置。我们拥有关于在内部使用过程中保持人工智能安全的最新技术见解的研究人员；与行业和其他前沿人工智能组织建立现有关系，并提供一个中立的第三方之家，共同建立最佳实践。这一角色直接实现了首相于2026年9月所做的承诺，即英国将致力于确保最强大的人工智能系统在安全且适当的透明度下开发的原则和标准。这直接基于我们2026年10月发布的出版物《迈向前沿人工智能开发的技术最佳实践》，阐述了我们的最紧迫优先事项的观点。
这是一个新的、正在成长的团队，具有高远的志向。这是AISI的一个跨部门倡议，汇集了AISI研究团队、以及我们的通讯、国际和战略团队的研究人员和工程师。我们现在正在建立一个专门的核心团队，并期待随着工作的进展而增长。
<h3>为什么这个团队很重要：</h3>
AISI独特地处于塑造此指导的位置，而该团队就是这样做的：
<ul>
<li>技术专长：AISI大约有10个技术和研究团队，其专业知识可以借鉴；这是在非前沿人工智能公司外从事人工智能风险工作的最大团队之一。AISI在发布前评估了55多个前沿模型，并在安全案例、可监测性和控制等方面发表了研究。</li>
</ul>
<ul>
<li>政府内的研究机构：我们直接与领先的AI开发者、其他政府和第三方评估者合作，并将他们聚集在没有任何一家公司会自行设定的实践周围。英国AISI也是……</li>
</ul>
</div>
</div>

🈯 翻译这份 JD · 红灯 0 / 黄灯 0 / 绿灯 0点开 ▾

没扫到常见信号词。这份 JD 直白得不常见——面试里照常确认节奏、汇报线和薪酬范围就好。

没扫到常见信号词——这份 JD 直白得不常见。照常确认节奏、汇报线和薪酬范围。

词典只标"常见信号"，不对公司下结论——红灯 = 面试里该当面确认的问题。想自己翻别的 JD：JD 翻译器 →

🏢

公司 · Company

Aisi

—

London来源：公开招聘平台

⚑ 岗位信息收录自公开招聘页（公开招聘平台），以雇主官网为准