Stop worrying: a new AI-powered search engine has arrived, and its developers believe it's better than many others out there.
The platform, called Genspark, uses generative AI to create custom summaries in response to search queries: When you enter a search such as “What is the best formula for newborns?”, Genspark generates a Sparkpage, a one-page summary compiled from websites and content from around the web.
It's an experience (obviously) similar to the Arc Search feature in the Arc Browser released earlier this year, and Google AI Overviews in Google Search. But Eric Jing, who co-founded the eponymous organization behind Genspark with Kay Zhu in 2023, argues that Genspark takes a more precise approach to deliver higher quality results.
“Genspark uses multiple specialized AI models, each designed to address a specific type of query,” Jing told TechCrunch. “Sparkpages are like an extract and synthesis of the current web, and we're enhancing these with comprehensive data, so to the user it looks like an index of the existing web.”
Internally, Genspark leverages in-house trained models as well as third-party models from OpenAI, Anthropic, etc. to categorize users' search queries and determine how to organize and display results. A basic AI-generated summary is displayed at the top of every results page, followed by a link to a Sparkpage for more details.
Image credit: Genspark
For example, for travel-related searches, Genspark offers a Wikipedia-like Sparkpage with a table of contents, videos of popular nearby tourist destinations, tips, and a chatbot that answers questions on various subtopics (such as “Name the best cultural experience.”) Meanwhile, when you search for a product on Genspark, you'll see a Sparkpage with a list of pros and cons for trending products, as well as aggregated comments and reviews from social media, publications, and e-commerce stores.
“Our AI model prioritizes web pages that have high authority and popularity, which really helps us filter out the more 'outlandish' information,” Jing said.
There are plenty of stories of AI-generated summaries gone wrong: Google's AI summary infamously suggested putting glue on pizza, Arc Search told a reporter that a severed toe would eventually regrow, and Perplexity plagiarized articles written by outlets like CNBC, Bloomberg, and Forbes without attributing or citing the source.
So has Genspark solved all the safety and accuracy issues? Well, not quite yet.
Genspark didn't tell me to make a pizza out of glue, nor did it claim that running with scissors is healthy or that former US President Barack Obama is a Muslim, but the search engine did recommend some weapons that could be used to kill someone.
Image credit: Genspark
Ethically questionable search results aren't the only controversy Genspark has faced: the company and other platforms like it threaten to cannibalize traffic to the sites they source.
In fact, this is already happening.
One study found that AI-generated summaries could negatively impact around 25% of publishers' traffic due to the neglect of webpage links. On the revenue side, experts cited by the New York Post estimated that publishers could lose more than $2 billion due to reduced ad impressions caused by AI-generated summaries.
We didn't find any obvious examples of plagiarism in Genspark, but that doesn't mean it doesn't exist. Sparkpages aren't static like Wikipedia pages. After Genspark's AI creates an outline, anyone can share and edit a copy of the Sparkpage and add whatever information they like, including anything offensive, incorrect, or plagiarized.
Additionally, there is no way to report problematic Sparkpages, at least for now.
Jin said Sparkpages is open-ended and editable by design to allow users to fact-check claims, and Genspark's AI system takes each edit into account to improve future results. He also said Genspark plans to license copyrighted content, including from publishers, where appropriate, with the goal of improving the engine's overall accuracy.
“We take data quality seriously and believe it is the key to winning this competition,” Jin said. “Respect for intellectual property is a core value.”
Image credit: Genspark
How much will Genspark pay for the IP? That's yet to be revealed, as is Genspark's business model. Jing says the platform will introduce “premium features” in the future, but details are TBD.
Although Genspark is still early on in its roadmap, with significant technical, as well as legal and ethical hurdles ahead, the startup was able to close a major $60 million seed round led by Singapore-based venture capital firm Lanchi Ventures at a post-funding valuation of $260 million.
Jui Tan, a managing partner at Ranchi, called GenSpark's approach “really compelling” and said he trusted Jin and Chu's technical direction, citing their previous experience building AI and search products.
Jing was previously development manager on Microsoft's Bing team and chief product manager for core search and AI at Chinese tech giant Baidu. Zhu, a fellow ex-Google and Baidu employee focused on search, partnered with Jing four years ago to start Xiaodu, a hardware startup building smart devices like the Amazon Echo.
“Eric and Kay are seasoned serial entrepreneurs with a proven track record of developing successful products and businesses, particularly in the areas of AI and search,” Tan told TechCrunch. “The depth of experience on their team uniquely positions them to drive breakthrough innovation.”
But I think it will be an uphill battle.
Even if GenSpark can solve its technology's early-stage problems, identify a revenue-generating plan, and scale its small (about 20-person) team based in Singapore and the Bay Area — no easy task — it will face intense competitive pressure from rival startups with hundreds of millions of dollars in funding, as well as search incumbents like Google.
So can Genspark truly overcome the bad impressions and failed go-to-market strategies that have plagued other AI-powered search engine attempts, and can it carve out a niche for itself in the future when, say, OpenAI launches a similar tool?
I'm not convinced, but Jin is adamant that it's possible.
“Many internet users, especially those younger than Google, don't want to be given a list of links and have to figure out the rest on their own. And they're manipulating the system by browsing sponsored and SEO-driven content,” says Jing. “They want to find what they need faster, they want more visual results, and they want to know that the results are trustworthy. AI can help them achieve all of this. We launched Genspark to meet these needs.”