The AI Beat: Why does OpenAI need a search engine? The AI Beat: Why does OpenAI need a search engine?
Join us in returning to NYC on June 5th to collaborate with executive leaders in exploring comprehensive methods for auditing AI models regarding bias,... The AI Beat: Why does OpenAI need a search engine?

Join us in returning to NYC on June 5th to collaborate with executive leaders in exploring comprehensive methods for auditing AI models regarding bias, performance, and ethical compliance across diverse organizations. Find out how you can attend here.

If the reports and rumors are to be believed, OpenAI, maker of ChatGPT and the company that is arguably most responsible for ushering in the generative AI boom (in terms of investment dollars and product offerings) in Silicon Valley and worldwide, is preparing to launch a new AI-powered search engine as early as this coming Monday, May 13.

But as regular users and paying subscribers of its flagship product ChatGPT will know, you can already get AI-powered searches directly through ChatGPT today.

Simply prompt/ask the chatbot (using GPT-4) to search the web for some topic or information, and it will automatically tap into Bing Search (from OpenAI backer and ally Microsoft) and present the information as well as links and citations.

Screenshot showing ChatGPT’s current search functionality with example search. Credit: OpenAI/Screenshot by author

Coming Soon: The AI Audit with UiPath

The latest stop on The AI Impact Tour is almost here!

VB Event

The AI Impact Tour: The AI Audit

Join us as we return to NYC on June 5th to engage with top executive leaders, delving into strategies for auditing AI models to ensure fairness, optimal performance, and ethical compliance across diverse organizations. Secure your attendance for this exclusive invite-only event.

Request an invite

Join us for an exclusive live event in NYC on June 5th as we delve into the fascinating world of AI Auditing at our latest event with our partners UiPath .

Discover strategies for examining AI models with fairness, performance, and ethical compliance.

You won’t want to miss out on meeting other industry leaders. Learn more:

How would OpenAI search engine be different than ChatGPT + Bing?

How could a separate or distinct OpenAI search engine possibly improve on this experience — and improve enough to warrant investing in the development and launch of something new?

Without seeing what OpenAI has planned, it’s hard to cast judgement or come to any firm conclusions about why a separate generative AI search experience would be more beneficial to users than the current offerings — or how it would reel in more potential users.

Indeed, there are already other AI-powered search engines out there, namely Perplexity and Google’s Search Generative Experience (SGE), the latter an optional ad on that users can toggle on when signed into their personalized Google account/profile.

But perhaps the very existence of these services has motivated OpenAI to launch a rival.

Competition ignites a fire

After all, Perplexity has received multi millions in funding for its offering — which lets users select different open source and proprietary AI models to power their searches — and is also quickly gaining traction among enterprises with a private, internal data search function.

With OpenAI also intently courting and winning over enterprise customers, it perhaps seems natural to them to have a search offering since thats what some enterprise customers are clamoring for — even though one is kind of already baked into ChatGPT.

ChatGPT is for conversation, analysis and work vs. OpenAI search for research?

Perhaps ChatGPT is actually too good, too all encompassing, and OpenAI sees an opportunity to atomize it slightly into some separate services and products such as a search-forward one.

Maybe the company has done research or has the perception — correct or not — that users view ChatGPT as a conversational assistant and content creation tool more than a research one, and wants to offer clear and distinct products for both use cases.

If so, it’s an interesting tack from a company that bundled its image generating AI model, DALL-E 3, directly into ChatGPT.

More is more?

Yet OpenAI has clearly also supported the “more is more” ideology with its offerings in recent months, and that is kind of the driving philosophy of its customized GPT Builder and GPT Store, as well as its Model Spec release showing that it is contemplating allowing users of its application programming interface (API) to even generate “not-safe-for-work” NSFW/explicit content (even pornography), which it heretofore sought to suppress and prevent users from generating with its products.

OpenAI may be taking the approach of Apple under Tim Cook — having multiple different products similar in functionality and use cases, but with slightly different configurations for different user groups and different price points.

Apple iPad models circa mid-2024. Credit: Apple Inc./Screenshot by author

There’s some advantage to that style of business: whatever the individual customer needs, no matter how unique, the message to them is “we have an offering among our multiple myriad ones that can hopefully fit you.”

But it’s also a dangerous road to take in that it can quickly overwhelm users with an overabundance of choice and paralyze them, slowing down their decision-making process or turning them away from the offerings entirely in favor of a simpler, cleaner, more “one-sized fits all” option from a rival.

Regardless, we’ll likely find out soon what OpenAI’s vision for search is beyond the current ChatGPT and Bing integration, and from there, we can begin to use it and analyze it.

Is it better than Google or Perplexity? What are its advantages and disadvantages over those other offerings? Who is it geared toward? Why would someone use it instead of the others? And what impact, if any is it likely to have on the search market but also the wider market of sources that feed into search.

That latter question is of course of special concern to publishers like us at VentureBeat, who rely heavily on Google Search referrals to get visitors and readers to our articles, as well as the accompanying advertising revenue we can generate from showing users ads directly on our site.

Publisher complaints and deals with the devil

Google has been criticized increasingly in recent months and years for robbing publishers of ad dollars by keeping visitors on Google Search Engine Results Pages (SERP), and for adjusting its algorithms in ways that seem to disadvantage smaller independent publishers, who are now competing with a tidal wave of AI generated content that is in many cases, directly ripping off their own work (or indirectly benefitting from it given that Google’s AI products and the larger corpus of generative AI large language models [LLMS] were trained on lots of publicly posted web content).

Will OpenAI’s presumed search product face the same line of criticism by providing users with information extracted from source articles without incentivizing users to actually visit those sources and provide them with traffic and ad impressions?

OpenAI has already smartly sought to head off some of this criticism by inking licensing deals directly with publishers including the Associated Press, Politico and Business Insider’s publisher Axel Springer, and more recently, the Financial Times — in which it plans to serve up their content or summaries of it as responses in ChatGPT along with links back to the sources.

Generative generation

Nonetheless, whatever happens, it’s clear the era of outbound link-ranked web search ushered in by Google more than 25 years ago is rapidly changing and, I would wager, coming to an end.

And it makes sense that it would at this time in particular. 25 years is a long time in tech and even human history, from a certain vantage point — the span of a generation. Now a new generation is emerging: a generative generation, to mince words a bit.

Maybe this, more than anything is why OpenAI seems poised to strike search next.

In an era when software ‘bots can do more than crawl and index the web — they can now read, summarize, translate, analyze, and interpret it — why should we humans be searching the same old way we did back in 1997?

The decline of Google’s perceived quality leaves an opening

Also, to be fair, Google’s search experience is widely perceived as having been degraded by the company itself in recent years as the top of the SERP has been increasingly filled with sponsored results placements and info boxes (it calls them “knowledge panels”) of varying accuracy and helpfulness rather than a link back to the information source the user is seeking.

The company has even been been sued by the U.S. government for antitrust over this very tendency (as has Amazon).

Which brings up the real question behind all this: what is search used for? Are people trying problem solve, have their questions answered as fast as possible? Or are they looking for underlying source material to help them answer it on their own?

There’s likely space for multiple approaches. Yet in an era that has been shaped by Google’s overwhelming search dominance (one that is quickly being challenged by AI) it’s hard to imagine a scenario in which one or a few companies — OpenAI, Google itself, Perplexity, Bing, some other startup — doesn’t emerge as the leader of the next era of search, and possibly, the web itself.

I don’t know if we can answer now why OpenAI thinks it needs a search engine, but to come back to that initial question, I’d hazard that the answer is: because it can build one, and maybe even a better one than what we have now. At least, that’s presumably the hope of Altman and co.

That’s all for this week! Have a great weekend. TGIF. Peace out.

Source link