内容来源：https://aibusiness.com/generative-ai/meta-s-muse-image-might-what-smbs-need

内容总结：

谷歌云赞助报道 | 企业选择首个生成式AI应用场景指南

Meta发布AI图像模型，助力中小企业广告创作

社交媒体巨头Meta近日推出其首个AI图像生成模型Muse Image，该工具可帮助中小企业在Facebook和Instagram等平台上创建具有竞争力的广告内容，使其广告效果堪比大型品牌和企业的专业制作。

据悉，Meta继今年4月推出Muse Spark多模态模型家族后，于7月7日正式发布由Meta超级智能实验室开发的Muse Image模型。该模型已嵌入Meta Advantage+自动化广告套件中，企业可通过自然语言指令，生成风格一致的品牌图像变体，并能修改背景等局部元素。

市场分析机构Constellation Research分析师莉兹·米利尔指出，Muse为中小企业提供了AI图像生成和编辑工具，帮助它们像大企业一样“搭乘AI浪潮”，尤其在图像编辑、混搭和生成方面获得效率提升。

不过米利尔也提醒，Meta此举背后有其商业考量。相比亚马逊广告等平台已提供的同类AI服务，Muse的技术能力并不领先。Meta真正的盈利点在于中小企业和普通消费者，而非那些早已建立自有AI内容创作流程的大型广告主。

“对于Meta而言，Muse是其推进AI能力布局的重要一步，”米利尔表示，“它能帮助中小企业获得个性化广告素材，同时有望提升Meta的广告收入，甚至可能超越谷歌。”

她进一步分析称，尽管Muse的能力可能永远比不上Midjourney或Adobe Firefly等主流图像生成模型，但对于中小企业而言，“它足以帮助它们像大机构和品牌一样获得定制化的营销资产”。

用户需注意数据安全风险

米利尔特别提醒用户保持审慎：需确保使用Muse Image生成的任何内容不被用于训练Meta的AI模型。她举例称，“如果你在Instagram上通过公开个人资料展示艺术作品，这些内容、作品和图像都有可能被用于模型训练。”

对于中小企业而言，数字安全和生成内容的商用合规性同样不容忽视。“小企业可能觉得这不是问题，但对于品牌方来说，确保生成的图像安全可靠、可直接用于商业用途，至关重要。”米利尔补充道。

中文翻译：

由谷歌云赞助
选择您的首个生成式AI应用场景
要开始使用生成式AI，首先应聚焦于能够改善人类获取信息体验的领域。

Meta的新AI图像生成模型使其平台用户能够创建具有竞争力的广告内容。

Meta表示，其首款AI图像生成模型使中小型企业能够创建与大型品牌和企业相媲美的广告及营销活动。

继今年4月推出Muse Spark多模态模型后，Meta于7月7日发布了由Meta超级智能实验室开发的Muse Image。该模型被嵌入Meta Advantage+自动化广告套件中，使企业能够利用该工具生成保持品牌一致性的图像变体。这款新模型具备智能代理功能，企业可通过输入自然语言指令，修改所生成图像的局部内容（如背景）。

对于中小型企业而言，Muse的发布使其能够使用AI驱动的图像生成与编辑工具，从而在Facebook和Instagram等Meta平台上提升运营效能。

“这帮助中小企业像大型企业一样乘上AI生成浪潮，尤其在编辑、重组和生成环节——这都归功于AI技术。”Constellation Research分析师莉兹·米勒表示。

米勒指出，尽管Muse为中小企业带来机遇，但Meta另有考量——尤其因为Muse的能力仍落后于亚马逊广告等提供类似AI服务的平台。

“Meta的盈利点可能不在于那些在Muse发布前就已建立成熟内容创作和AI生成流程的广告商与代理机构，”米勒分析道，“真正的收益将来自中小企业与普通用户，依托于Meta自主开发的AI模型。”

她补充道，Muse是Meta推进AI产品与能力升级的途径。该模型既能满足中小企业的实际需求，又能提升Meta的广告收入——其规模有望超越谷歌。

因此，尽管Muse可能永远无法达到Midjourney或Adobe Firefly等热门图像生成模型的水平，但这或许并不必要。米勒表示：“对于中小企业而言，它足以支撑其像大型机构与品牌一样获取‘个性化营销素材’。”

然而，用户需谨慎行事：必须确保使用Muse Image生成的任何内容都不会被用于训练Meta开发的任何模型。

“例如，若您作为艺术家通过公开个人主页在Instagram展示作品，这些内容、艺术品和图像都可能被用于模型训练。”米勒提醒道。她进一步强调，用户还应关注Muse生成内容的安全性问题。

“对小型企业而言这或许不是问题……但对品牌来说，数字安全以及确保生成的图像安全合规、可用于商业用途至关重要。”她总结道。

英文来源：

Sponsored by Google Cloud
Choosing Your First Generative AI Use Cases
To get started with generative AI, first focus on areas that can improve human experiences with information.
Meta’s new AI imaging model enables users to create competitive ad content on the social media giant’s platforms.
Meta said its first-ever AI imaging model enables small and mid-sized businesses to create ads and campaigns comparable to those of larger brands and enterprises.
After initially introducing the Muse family of multimodal models in April with Muse Spark, Meta released Muse Image, developed by the Meta Superintelligence Lab, on July 7. It is embedded in the Meta Advantage+ automated ad suite. This enables businesses to use the tool to generate brand-consistent image variations. The new model functions as an agent, and businesses can modify parts of the image it produces, such as the background, by typing natural-language instructions.
For small and medium-sized businesses, the release of Muse provides access to an AI-powered image-generation and editing tool that can boost their performance and efficiency on Meta’s platforms, such as Facebook and Instagram.
“This helps SMBs ride the wave of AI generation like larger businesses have, especially around editing, remixing and generation, thanks to AI,” said Liz Millier, an analyst at Constellation Research.
While Muse provides opportunities for SMBs, Meta has an ulterior motive, especially since Muse's capabilities lag those of other platforms, such as Amazon Ads, which offers similar AI-powered services, Miller said.
“The money maker for Meta may not be with advertisers and agencies who have likely already established their content creation and AI generation processes and flows before this later launch,” Miller said. “The money here is with the SMBs and the consumer, with an AI model that is developed and owned by Meta.”
She added that Muse is a way for Meta to advance its AI offerings and capabilities. Muse could deliver the results SMBs need, while boosting Meta’s ad spend, which could potentially surpass Google's.
Therefore, while Muse will likely never be as capable as popular image-generating models such as Midjourney or Adobe Firefly, it may not need to be. It might be just good enough for an SMB to get “personalized campaign assets just like the big agencies and brands,” Miller continued.
However, users need to exercise due diligence; it will be up to them to ensure that whatever is created with Muse Image is not used to train any Meta-developed model.
“If you are an artist showcasing work on Instagram via a public profile page, for example, that content, those artworks, those images are all potentially being used for training,” Miller said. She added that users should also ask questions about Muse-powered generations and safety.
“For a small business, this may simply not be an issue … but for brands, digital safety and the assurance that the images being generated are safe and ready for commercial use is massively important,” she added.