ChatGPT的高级语音模式上线：中文一开口，就暴露了歪果仁身份
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;"><span style="color: black;">设备</span>之心<span style="color: black;">报告</span></span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><strong style="color: blue;"><span style="color: black;">编辑：蛋酱、小舟</span></strong></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">OpenAI 的「Her」<span style="color: black;">最终</span>向部分人群开放了。</span></span></p>
<div style="color: black; text-align: left; margin-bottom: 10px;"><img src="https://p3-sign.toutiaoimg.com/tos-cn-i-axegupay5k/0e1a32772e0a4aff9c72202f37ec426c~noop.image?_iz=58558&from=article.pc_detail&lk3s=953192f4&x-expires=1724858766&x-signature=NcCVM97iBezt%2B%2F0bt6DWS6CACUQ%3D" style="width: 50%; margin-bottom: 20px;"></div>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">今年 5 月，OpenAI 在「春季新品发布会」上搬出了新一代旗舰生成模型 GPT-4o、桌面 App，并展示了一系列新能力。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;"><span style="color: black;">此刻</span>，OpenAI 宣布向一小部分 ChatGPT Plus 用户开放 ChatGPT 的高级语音模式，让用户首次<span style="color: black;">得到</span> GPT-4o 的超现实音频响应。这部分用户将在 ChatGPT 应用程序中收到提醒，并收到一封电子邮件，其中<span style="color: black;">包括</span><span style="color: black;">相关</span><span style="color: black;">怎样</span><span style="color: black;">运用</span>该应用程序的说明。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">「自从<span style="color: black;">咱们</span>首次演示先进的语音模式<span style="color: black;">败兴</span>，<span style="color: black;">咱们</span><span style="color: black;">始终</span>致力于加强语音对话的安全性和质量，准备将这项前沿技术带给数百万人。」OpenAI <span style="color: black;">暗示</span>，该功能将在 2024 年秋季逐步向所有 Plus 用户推出。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;"><span style="color: black;">有些</span>用户<span style="color: black;">已然</span>晒出了高级语音模式的<span style="color: black;">运用</span>效果：</span></span></p>
<div style="color: black; text-align: left; margin-bottom: 10px;">
重播
<div style="color: black; text-align: left; margin-bottom: 10px;">播放</div>
<span style="color: black;">00:00</span>
<span style="color: black;">/</span>
<span style="color: black;">00:00</span>
<span style="color: black;">直播</span>
<div style="color: black; text-align: left; margin-bottom: 10px;">
<div style="color: black; text-align: left; margin-bottom: 10px;">00:00</div>
</div>
<div style="color: black; text-align: left; margin-bottom: 10px;">进入全屏</div>
<div style="color: black; text-align: left; margin-bottom: 10px;">50</div>
<div style="color: black; text-align: left; margin-bottom: 10px;">点击按住可拖动视频</div>
</div>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">源自</span>：</p>https://x.com/tsarnick/status/1818402307115241608
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">当你和 ChatGPT 讲段子时，Ta <span style="color: black;">能够</span><span style="color: black;">供给</span><span style="color: black;">有些</span>笑声<span style="color: black;">陪同</span>：</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">源自</span>：</p>https://x.com/yoimnotkesku/status/1818406786077970663
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;"><span style="color: black;">运用</span> ChatGPT 的高级语音模式，「Her」<span style="color: black;">能够</span>在讲故事的<span style="color: black;">同期</span>创建背景音乐，并且适用于多种语言。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">源自</span>：</p>https://x.com/yoimnotkesku/status/1818415019349901354
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">法语、西班牙语和乌尔都语<span style="color: black;">亦</span>都<span style="color: black;">能够</span>：</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">源自</span>：</p>https://x.com/yoimnotkesku/status/1818424494106853438
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">但中文表达不太地道，仿佛一个正在学习中文的「歪果仁」：</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">源自</span>：</p>https://x.com/yoimnotkesku/status/1818446895083139170
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">听完的人都懵了：</span></span></p>
<div style="color: black; text-align: left; margin-bottom: 10px;"><img src="https://p3-sign.toutiaoimg.com/tos-cn-i-6w9my0ksvp/70b2f7710fdc4b6296e9a1d8594f796e~noop.image?_iz=58558&from=article.pc_detail&lk3s=953192f4&x-expires=1724858766&x-signature=xcAB0hxda%2BM%2BDswOR6u0wGvcGik%3D" style="width: 50%; margin-bottom: 20px;"></div>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">而口音问题不只出<span style="color: black;">此刻</span>中文，<span style="color: black;">据述</span>德语<span style="color: black;">亦</span><span style="color: black;">同样</span>：</span></span></p>
<div style="color: black; text-align: left; margin-bottom: 10px;"><img src="https://p3-sign.toutiaoimg.com/tos-cn-i-6w9my0ksvp/16841c79f62e4175a2b0515086184851~noop.image?_iz=58558&from=article.pc_detail&lk3s=953192f4&x-expires=1724858766&x-signature=O8pkxcAmy5l%2BdjfMPueb1Lk3uAM%3D" style="width: 50%; margin-bottom: 20px;"></div>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">源自</span>：</p>https://x.com/yoimnotkesku/status/1818445235606671670
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">最后，讲段绕口令吧：</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;">源自</span>：</p>https://x.com/yoimnotkesku/status/1818427991514337695
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">OpenAI <span style="color: black;">暗示</span>高级语音模式与 ChatGPT <span style="color: black;">日前</span><span style="color: black;">供给</span>的语音模式有所<span style="color: black;">区别</span>。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">ChatGPT 的旧语音模式<span style="color: black;">处理</span><span style="color: black;">方法</span><span style="color: black;">运用</span>了三种独立的模型：一个模型将语音转换为文本，GPT-4 负责处理提示（prompt），第三个模型则负责将 ChatGPT 的文本转换为语音。而 GPT-4o 是多模态的，能够在<span style="color: black;">无</span>辅助模型的<span style="color: black;">帮忙</span>下处理这些任务，从而<span style="color: black;">明显</span>降低对话延迟。OpenAI 还<span style="color: black;">暗示</span> GPT-4o <span style="color: black;">能够</span>感知用户声音中的<span style="color: black;">心情</span>语调，<span style="color: black;">包含</span><span style="color: black;">哀痛</span>、兴奋等等。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">今年 5 月，OpenAI 首次展示了 GPT-4o 的语音功能，「她」的反应速度、与真人声音的惊人<span style="color: black;">类似</span>度震惊了观众 —— 问题就出在这儿。</span></span></p>
<div style="color: black; text-align: left; margin-bottom: 10px;"><img src="https://p3-sign.toutiaoimg.com/tos-cn-i-6w9my0ksvp/82fac6c0b0b94ec389899f1db8f88eb1~noop.image?_iz=58558&from=article.pc_detail&lk3s=953192f4&x-expires=1724858766&x-signature=3GZCSq10Vky8pUSuz74z%2BK9XueM%3D" style="width: 50%; margin-bottom: 20px;"></div>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">这个名叫 「Sky」 的声音酷似电影《Her》中人工助手的扮演者斯嘉丽・约翰逊（Scarlett Johansson）。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">在 OpenAI 演示之后不久，约翰逊说她曾拒绝 OpenAI CEO 山姆・奥特曼关于<span style="color: black;">运用</span>她的声音的多次请求，在看到 GPT-4o 的演示之后，她聘请了法律顾问为自己的声音辩护。OpenAI 否认<span style="color: black;">运用</span>了斯嘉丽・约翰逊的声音，但<span style="color: black;">亦</span>删除了演示中的声音。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">6 月，OpenAI <span style="color: black;">暗示</span>将推迟发布高级语音模式，以改进其安全<span style="color: black;">办法</span>。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">漫长的等待后，「Her」总算与<span style="color: black;">大众</span>见面了。OpenAI <span style="color: black;">暗示</span>，此次推出的高级语音模式将仅限于 ChatGPT 与付费配音演员合作，制作了四种预设语音：Juniper、Breeze、Cove 和 Ember。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">值得<span style="color: black;">重视</span>的是，输出的声音有且<span style="color: black;">仅有</span>这四种 —— OpenAI 5 月份的演示中展示的 Sky 语音已<span style="color: black;">再也不</span>适用于 ChatGPT。OpenAI 发言人 Lindsay McCallum <span style="color: black;">暗示</span>：「ChatGPT <span style="color: black;">不可</span>冒用他人的声音，<span style="color: black;">包含</span>个人和公众<span style="color: black;">名人</span>的声音，并且会阻止与这些预设声音之一<span style="color: black;">区别</span>的输出。」</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">这种设置的初衷是避免 Deepfake 争议。今年 1 月，人工智能初创<span style="color: black;">机构</span> ElevenLabs 的语音克隆技术被用来冒充美国总统拜登，<span style="color: black;">诈骗</span>了新罕布什尔州的初选选民，<span style="color: black;">诱发</span>了不小的争议。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">OpenAI 还<span style="color: black;">暗示</span>，<span style="color: black;">已然</span>引入了新的过滤器来阻止某些生成音乐或其他受版权<span style="color: black;">守护</span>音频的请求。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">去年，<span style="color: black;">非常多</span>图像生成、音乐生成的 AI <span style="color: black;">机构</span>因侵犯版权而陷入了法律纠纷，尤其是<span style="color: black;">爱好</span>打官司的唱片<span style="color: black;">机构</span>，<span style="color: black;">已然</span>起诉过人工智能音频生成器 Suno 和 Udio。而 GPT-4o <span style="color: black;">这般</span>的音频模型则让<span style="color: black;">能够</span>提出投诉的<span style="color: black;">机构</span><span style="color: black;">增多</span>了一个全新的类别。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;"><span style="color: black;">据述</span>，OpenAI 与 45 种语言的 100 多名<span style="color: black;">外边</span>「红队」成员<span style="color: black;">一块</span>测试了 GPT-4o 的语音功能。而这些关键信息，将在 8 月份一份关于 GPT-4o 的功能、局限性和安全<span style="color: black;">评定</span>报告中有更<span style="color: black;">仔细</span>的<span style="color: black;">颁布</span>。</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">参考链接：</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">https://twitter.com/OpenAI/status/1818353580279316863</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">https://www.theverge.com/2024/7/30/24209650/openai-chatgpt-advanced-voice-mode</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">https://www.reuters.com/technology/openai-starts-roll-out-advanced-voice-mode-some-chatgpt-plus-users-2024-07-30/</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">https://www.bloomberg.com/news/articles/2024-07-30/openai-begins-rolling-out-voice-assistant-after-safety-related-delay?srnd=phx-technology</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">https://techcrunch.com/2024/07/30/openai-releases-chatgpts-super-realistic-voice-feature/</span></span></p>
<p style="font-size: 16px; color: black; line-height: 40px; text-align: left; margin-bottom: 15px;"><span style="color: black;"><span style="color: black;">https://www.theinformation.com/briefings/after-delay-openai-releases-ai-voice-assistant</span></span></p>
论坛外链网http://www.fok120.com/ 对于这个问题，我有不同的看法...
页:
[1]