Search and AI have become inextricably entwined. On the recent INMA Silicon Valley Study Tour participants had the opportunity to hear and question executives from companies at 20 different stops, including Google and Andi, on what is coming to the field.
Both companies, among the 20+ stops on the five-day tour attended by 18 international news media leaders, are actively trying to weed out spam and disinformation from search results. Both make the case that AI can be used to help newsrooms do more with less and to help contextualise information to build trust in uncertain times. And both believe news organisations make worthy partners.
Google: Pinpoint and Search Generative Experiences
At Google, participants engaged with speakers who look at search, AI, and news from different positions, including:
Using Google’s Pinpoint, which uses speech-to-text and OCR technologies that allow quick searches through documents, images, and audio recordings — as tools for telling stories with data — for example, how shopping patterns changed during the pandemic.
How Google’s different iterations of search — Discover, News, and Search — which use AI to contextualise searches and provide better answers, have different functions in the search ecosystem.
News organisations tend to rank high in categories Google uses to evaluate search quality: expertise, authority, trustworthiness, and transparency (EATT). Relatedly, the trajectory for Google product development, including Search Generative Experiences (SGEs, which uses AI to assist in finding key points on Web pages) is to enhance rather than replace journalism, participants were assured.
Using AI in search is still relatively new, one speaker said: “We don’t know where the industry is going.” But the general challenge, they said, is being open to change after accomplishing mastery at current tasks.
Startup Andi creates search for the next generation
Andi, another stop on the tour, has both disruptive technology and a disruptive business paradigm. Nick Chan, COO, described Andi as purpose- as opposed to product-led. The company hopes to completely transform the search landscape by challenging the dominant search platforms.
Andi is trying to surface great content using generative AI to not only provide search results in visual, as well as textual formats, but to direct users to the links from which the results are derived. And it delivers results faster than traditional search engines.
“We are creating search for the next generation … to build something 100 times better,” said Ang Hoover, CEO and co-founder of Andi. “It will be visual, factual, and accurate,” as well as “conversational.”
She contrasted that with TikTok, an informational tool and a source of information for the next generation — where one in five videos contain false information, Hoover said. That begs the question: “How do we engage the Gen Z audience and still maintain quality?”
One of the things Hoover said she learned at Y Combinator, the U.S. technology startup accelerator backing this AI-based search startup, is that writing code and talking to users are the two things that matter. Sourcing matters, too: “We do base results on what is canonical content … but we aim to uplift quality content that is not made up and to send users to high-quality content producers.”
“From our side,” Chan said, “there is no censorship either on the right or the left.” By using a semantic model, in which the search engine understands the question, Andi endeavours to serve the content it believes the user is searching for.
“We believe that users have a right to rank sources for themselves,” Hoover said. “Our job is to display the highest quality content, and to push the decision-making process back to the user.”