Chinese startup DeepSeek AI and its open-source language designs took over the news cycle today. Besides being similar to models like Anthropic's Claude and OpenAI's o1, the models have raised numerous issues about data personal privacy, security, and Chinese-government-enforced censorship within their training.


AI search platform Perplexity and AI assistant You.com have actually discovered a way around that, albeit with some restrictions.


Also: I evaluated DeepSeek's R1 and V3 coding abilities - and we're not all doomed (yet)


On Monday, Perplexity posted on X that it now hosts DeepSeek R1. The free strategy gives users three Pro-level questions per day, which you might use with R1, however you'll need the $20 each month Pro plan to gain access to it more than that.


DeepSeek R1 is now readily available on Perplexity to support deep web research. There's a brand-new Pro Search reasoning mode selector, together with OpenAI o1, with transparent chain of believed into model's reasoning. We're increasing the variety of day-to-day usages for both complimentary and paid as add more ... pic.twitter.com/KIJWpPPJVN


In another post, the business validated that it hosts DeepSeek "in US/EU information centers - your data never ever leaves Western servers," ensuring users that their data would be safe if using the open-source designs on Perplexity.


"None of your information goes to China," Perplexity CEO Aravind Srinivas repeated in a LinkedIn post.


Also: Apple researchers reveal the secret sauce behind DeepSeek AI


DeepSeek's AI assistant, powered by both its V3 and R1 models, is accessible by means of web browser or app-- but those need interaction with the company's China-based servers, which develops a security threat. Users who download R1 and run it in your area on their devices will avoid that issue, but still run into censorship of particular topics figured out by the Chinese government, as it's integrated in by default.


As part of offering R1, Perplexity claimed it got rid of at least some of the censorship built into the model. Srinivas posted a screenshot on X of question results that acknowledge the president of Taiwan.


However, when I asked R1 about Tiananmen Square using Perplexity, the design refused to respond to.


When I asked R1 if it is trained not to respond to certain questions identified by the Chinese government, it reacted that it's created to "concentrate on factual information" and "prevent political commentary," and that its training "highlights neutrality in global affairs" and "cultural level of sensitivity."


"We have removed the censorship weights on the design, so it should not act in this manner," said a Perplexity representative reacting to ZDNET's ask for comment, including that they were looking into the problem.


Also: What to understand about DeepSeek AI, from cost claims to information privacy


You.com uses both V3 and R1, similarly just through its Pro tier, which is $15 per month (marked down from the usual $20) and with no totally free queries. In addition to access to all the models You.com offers, the Pro strategy includes file uploads of as much as 25MB per query, a 64k maximum context window, and access to research and custom-made agents.


Bryan McCann, You.com cofounder and CTO, described in an email to ZDNET that users can access R1 and V3 by means of the platform in 3 ways, all of which utilize "an unmodified, open source version of the DeepSeek designs hosted entirely within the United States to make sure user privacy."


"The first, default method is to use these designs within the context of our proprietary trust layer. This provides the designs access to public web sources, a predisposition towards citing those sources, and a disposition to respect those sources while generating reactions," McCann continued. "The 2nd way is for users to turn off access to public web sources within their source controls or by using the models as part of Custom Agents. This choice enables users to check out the models' special capabilities and habits when not grounded in the general public web. The 3rd way is for users to check the limits of these designs as part of a Custom Agent by adding their own instructions, files, and sources."


Also: The very best open-source AI designs: All your free-to-use alternatives explained


McCann kept in mind that You.com compared DeepSeek models' actions based on whether it had access to web sources. "We observed that the models' reactions varied on numerous political topics, in some cases refusing to address on specific concerns when public web sources were not consisted of," he explains.