e-biz*solution
What is your confidence in the ‘Perplexity Search AI’ that threatens Google?… “Our strength is that we can load external data”
Digital Daily Publication date 2024-11-04 18:35:17
SK AI Summit 2024‘ through live streaming on the 4th, is giving a presentation on the topic of ‘Creating a more exploratory and efficient world with Perplexity’s AI search.'”/>
Perplexity CEO Aravind Srinivas, who participated in the ‘SK AI Summit 2024’ through live streaming on the 4th, is giving a presentation on the topic of ‘Creating a more exploratory and efficient world with Perplexity’s AI search.’
[디지털데일리 오병훈기자] As the Spring and Autumn Warring States era in the generative artificial intelligence (AI) market has arrived, interest in generative AI services exclusively for companies is also growing. In line with this, the global AI search company Perplexity also appears to have chosen ‘enterprise AI‘ for companies as its new business, which provides a new perspective by synthesizing all internal and external information of the company.
On the 4th, Perplexity CEO Aravins Srinivas made a presentation on the theme of ‘Creating a more exploratory and efficient world with Perplexity’s AI search’ at the ‘SK AI Summit 2024’ held at Samsung COEX and said, “Sometime next month. The new AI search model to be released can load not only its own data but also external data into Perplexity. “As a result, it will be a search model that can conduct exclusive data research related to verticals,” he said.
CEO Srinivas, who attended the event via video call, touted the ability to search information by combining internal and external data as the greatest strength of the newly introduced corporate AI search service. The explanation is that although source data comes from internal and external sources, the AI model that synthesizes this can provide exclusive information that is not available in the market.
He said, “For example, in the stock market, it can be used to provide new implications by adding user personal data to external stock prices and corporate information,” and added, “Specific financial information based on data such as performance, price-earnings ratio (PER), and cash flow. “We can collect information,” he emphasized.
In addition, Perplexity also establishes a publisher compensation system to resolve information source conflicts. Due to the nature of generative AI, it is not free from copyright issues because it uses information such as news and blogs that exist on the web. In fact, several generative AI companies, including Perplexity in the United States, have experienced conflicts with media companies over copyright issues.
The card that Perplexity came up with to solve this problem is the publisher compensation system. This is a method of providing a portion of revenue to companies that provide source data to Perplexity.
CEO Srinivas said, “We are preparing a system to properly compensate publishers (source data providers),” and added, “For media companies, we plan to share profits through the publisher compensation program.”
Meanwhile, Perplexity is a search AI service established in 2022. A variety of models are being used, including OpenAI’s ‘ChatGPT’ series, Anthropic’s ‘Claude’, and Mistral’s ‘7B’. Based on the model, we focused on mitigating the hallucination problem and improving search quality through search-specific fine tuning. As a result, it was able to quickly grow into a search AI startup that attracts attention in the global generative AI market.
Lastly, he said, “The search market has been stagnant for the past decade,” and “Since the search AI market was newly opened, Perplexity has focused on providing users with simple search accessibility that excludes unnecessary information and advertisements.”
