内容来源：https://aibusiness.com/generative-ai/perplexity-getty-images-licensing-deal

内容总结：

人工智能搜索服务商Perplexity AI与视觉内容提供商盖蒂图像于10月30日宣布达成多年期内容授权协议，此举为AI行业内容合作模式带来新范式。根据协议，Perplexity将获得盖蒂图像编辑类及创意类图片库的调用权限，并在展示图片时添加来源链接与版权标识，引导用户通过正规渠道使用授权内容。

值得注意的是，此次合作区别于常规AI数据训练协议——Perplexity未被授予使用盖蒂图像内容训练其AI模型的权限，仅可通过类API方式将版权内容接入搜索结果，并始终跳转至盖蒂官方页面完成最终呈现。分析师阿伦·钱德拉塞卡兰指出，这标志着数据服务商正通过激活视觉知识产权开辟新型营收渠道。

华盛顿大学教授奇拉格·沙阿认为，该合作更接近传统搜索引擎的"内容指针"模式，相较于直接将授权内容用于模型训练，这种保留内容方核心价值的合作更具可持续性。尽管协议明确了版权归属，但专家表示，若AI平台未来自主生成廉价替代内容，仍可能对原创者构成潜在挑战。

目前Perplexity已与《洛杉矶时报》《财富》等媒体建立发布合作，但同时也面临包括Reddit在内的多起数据侵权诉讼。此次与盖蒂图像的创新合作模式，或将为行业建立"内容权属契约化"标准提供重要参考。

中文翻译：

由谷歌云赞助
选择首个生成式AI应用场景
开展生成式AI项目时，应优先关注能够优化人类信息交互体验的领域。

与其他协议不同，虽然AI搜索供应商现可获取视觉媒体公司的内容资源，但并未获得内容训练授权。Perplexity AI与盖蒂影像达成的全新多年期许可协议，标志着AI版权合作模式的显著转变。

盖蒂方面表示，该协议将允许Perplexity使用其新闻编辑类及创意类图像资源。此外，Perplexity将调整图片展示方式，包括标注来源链接以确保用户合法使用授权图像，并注明版权归属盖蒂影像。在10月30日宣布合作前，盖蒂影像已参与Perplexity的出版商计划长达一年。

除盖蒂外，Perplexity还与Adweek、甘尼特集团、《洛杉矶时报》、《财富》等媒体机构保持出版合作。尽管已建立多方合作，该公司仍面临多起数据盗用与版权侵权指控，近期更因数据抓取行为遭到Reddit起诉。

高德纳分析师阿伦·钱德拉塞卡指出，此次合作对双方均具有积极意义。"通过授权图像库实现视觉IP变现，是盖蒂影像在AI时代的重要布局。数据服务商与AI企业达成合作，将数据资产转化为新收入来源，这已成为行业趋势。"

值得关注的是，与同类协议不同，本次合作禁止Perplexity使用盖蒂图像训练AI模型，仅允许其在搜索结果中呈现盖蒂内容并设置原图链接。这种模式类似于API接口调用——用户查看图片时始终被引导至盖蒂官方网站。

西雅图华盛顿大学教授奇拉格·沙阿评价道："相较于那些授权内容用于模型训练的协议，本案更显公平。某些版权协议对内容提供商存在潜在风险，当AI公司认定内容价值衰减时，合作便难以为继。"他还指出："这种传统跳转模式类似谷歌早期做法，通过搜索引导用户访问源网站，而非直接复制内容，我认为这是更具可持续性的发展路径。"

钱德拉塞卡补充说明，Perplexity的图片来源标注机制体现了对"版权规范"的重视。对企业而言，该协议表明"内容权益不再只是优选项，而是正在成为合同刚需"。不过沙阿也提出，虽然协议规范了版权归属，但对与盖蒂直接签约的内容创作者会产生何种影响尚不明确。

沙阿进一步分析，尽管无法使用盖蒂图像训练模型，Perplexity仍可能生成用户更青睐的替代内容，特别是在自制内容成本低于授权图像的情况下。这种情形类似于沃尔玛同时销售自有品牌与第三方商品，未来Perplexity或许会在引导用户访问盖蒂的同时，自行产出同类图像资源。

相关推荐

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.
While the AI search vendor now has access to the visual media company's content, it has not been given permission to train its AI model on the content, which is different from other deals.
AI search vendor Perplexity AI and Getty Images' new multi-year licensing agreement marks a distinct change from previous AI licensing deals.
Getty said the agreement will enable Perplexity to access its editorial and creative imagery. In addition, Perplexity will make changes to how it displays images, including crediting Getty Images by adding a link to the source so that users can use the licensed image legally, Getty said.
Before the partnership deal, revealed on October 30, Getty had been part of Perplexity's publisher program for the last year.
Perplexity maintains publishing arrangements with other media companies and outlets, including Adweek, Gannett, The Los Angeles Times and Fortune.
Despite its relationship with other publishers, the vendor faces multiple allegations of data theft or copyright infringement. Most recently, Reddit sued the company for data scraping.
The deal between Getty and Perplexity is favorable for both organizations, said Arun Chandrasekaran, an analyst at Gartner.
"Getty is monetizing its visual IP in the AI era by licensing its image library," Chandrasekaran said. "This is a continuing trend of data providers inking partnerships with AI companies to monetize the data assets as a new revenue model."
However, the deal with Getty differs from similar setups in that Perplexity can't train its model on Getty Images' data. Instead, the agreement allows Perplexity to provide Getty's content in its search results, linking back to those images. Perplexity gains access to Getty Images, much like an API does. Whenever users want to view an image, they are still directed to the Getty Images website.
"I see this as more of a fair deal than some of the other deals where AI companies have made with the content providers, which essentially license their content to train the models," said Chirag Shah, a professor at the University of Washington in Seattle. He added that some licensing deals can be dangerous for content providers because, at some point, AI companies will find their content less valuable and they no longer need it.
"[The Getty-Perplexity deal] is more of a traditional pointer, the kind of thing that traditional Google has done, where you search and then you click on something that takes you to that content provider or that website, as opposed to Google reproducing it right here," Shah said. "That's why I feel like this is a more sustainable model."
The system in which Perplexity links to Getty shows that the vendor is committed to "proper image credit," Chandrasekaran said.
For business, this deal demonstrates that "content rights are no longer 'nice-to-have', but they are becoming contractual,' Chandrasekaran continued.
Although this contract addresses attribution, it's unclear what effect it will have on content creators who have a direct agreement with Getty Images, Shah said.
While Perplexity can’t train its models on Getty Images content, it can still produce similar content that users might prefer, especially if the content Perplexity generates is less expensive than the images it links to, Shah said.
That scenario is similar to a scenario in which Walmart offers an in-house brand but continues to sell products from other producers, Shah continued. Perplexity might reach a point where it points back to Getty Images but also generates its own similar images.
You May Also Like