披露设计准则Disclosure design guidelines

了解如何通过透明化语音体验的合成性质来构建和维护客户给予的信任。Learn how to build and maintain trust with customers by being transparent about the synthetic nature of your voice experience.

什么是公开?What is disclosure?

公开是一种让人们知道他们正在听一个合成的语音或与其互动的方式。Disclosure is a means of letting people know they're interacting with or listening to a voice that is synthetically generated.

为什么需要公开?Why is disclosure necessary?

公开计算机生成语音的合成来源的需要是近年来出现的。The need to disclose the synthetic origins of a computer-generated voice is relatively new. 过去,计算机生成的语音很明显,没有人会把它们误认为是真人。In the past, computer-generated voices were obviously that—no one would ever mistake them for a real person. 然而,合成语音的逼真度日益提高,人们越来越难将其与人类的声音区分开来。Every day, however, the realism of synthetic voices improves, and they become increasingly indistinguishable from human voices.

目标Goals

在设计合成语音体验时,需牢记以下原则:These are the principles to keep in mind when designing synthetic voice experiences:

强化信任Reinforce trust
设计的意图是让图灵测试不通过,但不降低体验。Design with the intention to fail the Turing Test without degrading the experience. 让用户了解到他们正与合成语音进行交互,同时让他们能够无缝参与体验。Let the users in on the fact that they're interacting with a synthetic voice while allowing them to engage seamlessly with the experience.

适应使用的上下文Adapt to context of use
了解用户何时、何地以及如何与合成语音进行交互,以便在适当的时间提供正确的公开类型。Understand when, where, and how your users will interact with the synthetic voice to provide the right type of disclosure at the right time.

设置明确的期望Set clear expectations
让用户能够轻松发现和理解代理的功能。Allow users to easily discover and understand the capabilities of the agent. 根据请求提供机会,以了解更多有关合成语音技术的详细信息。Offer opportunities to learn more about synthetic voice technology upon request.

接纳失败Embrace failure
将失败作为强化代理功能的契机。Use moments of failure to reinforce the capabilities of the agent.

如何使用本指南How to use this guide

本指南可帮助你确定哪种公开模式最适合你的合成语音体验。This guide helps you determine which disclosure patterns are best fit for your synthetic voice experience. 然后,我们提供有关如何以及何时使用它们的示例。We then offer examples of how and when to use them. 其中每个模式都旨在最大程度地面向用户提高合成语音的透明度,同时保持以人为本的设计。Each of these patterns is designed to maximize transparency with users about synthetic speech while staying true to human-centered design.

考虑到语音体验设计指南内容繁多,我们侧重于介绍:Considering the vast body of design guidance on voice experiences, we focus here specifically on:

  1. 公开评估:确定为合成语音体验推荐的公开类型的过程Disclosure assessment: A process to determine the type of disclosure recommended for your synthetic voice experience

  2. 公开方式:可应用于合成语音体验的公开模式的示例How to disclose: Examples of disclosure patterns that can be applied to your synthetic voice experience

  3. 公开时间:在整个用户体验之旅中进行公开的最佳时间When to disclose: Optimal moments to disclose throughout the user journey

公开评估Disclosure assessment

考虑用户对交互的期望以及他们体验语音时所处的上下文。Consider your users' expectations about an interaction and the context in which they will experience the voice. 如果上下文清楚地表明合成语音正在"说",则公开可能是少量且短暂的,甚至是没有必要的。If the context makes it clear that a synthetic voice is "speaking," disclosure may be minimal, momentary, or even unnecessary. 影响公开的上下文的主要类型包括角色类型、场景类型和公开级别。The main types of context that affect disclosure include persona type, scenario type, and level of exposure. 这也助于考虑谁会听到这些内容。It also helps to consider who might be listening.

了解上下文Understand context

使用此工作表来确定合成语音体验的上下文。Use this worksheet to determine the context of your synthetic voice experience. 在下一步中,你将应用此工作表来确定公开级别。You'll apply this in next step where you'll determine your disclosure level.

使用的上下文Context of use 潜在风险和挑战Potential Risks & Challenges
1.角色类型1. Persona type 如果满足以下任一条件,则你的角色属于“似人类角色”类别:If any of the following apply, your persona fits under the 'Human-like Persona' category:

  • 无论是否是虚构的表现形式,角色体现真实的人。Persona embodies a real human whether it's a fictitious representation or not. (例如,真人的照片或计算机生成的渲染)(e.g., photograph or a computer-generated rendering of a real person)

  • 合成语音基于辨识度高的真人(例如名人、政治人物)的语音The synthetic voice is based on the voice of a widely recognizable real person (e.g., celebrity, political figure)
你赋予角色的似人类表示形式越多,用户将其与真人联系起来的可能性就越大,或者会导致他们相信内容是真人说的而不是计算机生成的。The more human-like representations you give your persona, the more likely a user will associate it with a real person, or cause them to believe that the content is spoken by a real person rather than computer-generated.
2.场景类型2. Scenario type 如果满足以下任一条件,则语音体验属于“敏感”类别:If any of the following apply, your voice experience fits under the 'Sensitive' category:

  • 获取或显示用户的个人信息Obtains or displays personal information from the user

  • 广播时效性新闻/信息(例如紧急警报)Broadcasts time sensitive news/information (e.g., emergency alert)

  • 旨在帮助真人之间的交流(例如,阅读个人电子邮件/短信)Aims to help real people communicate with each other (e.g., reads personal emails/texts)

  • 提供医疗/保健援助Provides medical/health assistance
当主题涉及敏感、个人或紧急事件时,使用合成语音可能会让人觉得不恰当或不可信。The use of synthetic voice may not feel appropriate or trustworthy to the people using it when topics are related to sensitive, personal, or urgent matters. 他们还可能期望具有与真人相同级别的同理心和上下文感知。They may also expect the same level of empathy and contextual awareness as a real human.
3.公开级别3. Level of exposure 如果出现以下情况,你的语音体验最可能属于“高”类别:Your voice experience most likely fits under the 'High' category if:

  • 用户会经常或长时间听到合成语音或与之交互The user will hear or interact with the synthetic voice frequently or for a long duration of time
在建立长期关系时,保持透明度以及与用户之间建立信任的重要性甚至更高。The importance of being transparent and building trust with users is even higher when establishing long-term relationships.

确定公开级别Determine disclosure level

根据使用的上下文,利用下面的图表来确定合成语音体验需要高公开级别还是低公开级别。Use the following diagram to determine whether your synthetic voice experience requires high or low disclosure based on your context of use.

公开评估图表

参考文档Reference docs

后续步骤Next steps