Despite swirling rumors of an impending search engine launch, OpenAI, the creator of popular chatbots like ChatGPT and a recipient of substantial investment from Microsoft, officially debunked the speculation.
A spokesperson for OpenAI firmly stated to The Register that "contrary to reports, we are not launching a search product or GPT-5 on Monday," directly addressing rumors of a product update scheduled for early next week. CEO Sam Altman echoed this denial through social media postings.
This clarification likely comes as a relief to Microsoft, which has integrated OpenAI's ChatGPT into its Bing search engine. However, for Reuters, who published a report claiming the AI powerhouse planned to unveil a Google Search competitor this week, this serves as a moment to reflect on sourcing accuracy or an opportunity to double down on their reporting later.
While the timing - one day before Google's annual developer conference, Google I/O - seemed strategically advantageous for a new product debut, we're inclined to trust OpenAI's official statement.
OpenAI has only promised to "show some updates to ChatGPT and GPT-4." Yet, this commitment seems underwhelming to warrant global press briefings, social media posts, email "press holds," and livestream invitations. Speculation points to OpenAI potentially showcasing a voice-controlled assistant.
Altman's teasing of Monday's unveiling - "We've been working hard on some new things that we think people will love! Feels like magic to me" - sounds reminiscent of Apple's cryptic marketing tactics. So, the new offering may not be G-spotPT, or whatever OpenAI's eventually unveiled NSFW model is called.
Given the chatter about OpenAI forging partnerships with publishers, the AI company might demonstrate how its models can summarize up-to-the-minute news content in chatbot replies, a feature functionally overlapping with search engines.
Perhaps this is the correct way to conceptualize how machine learning models will integrate with the search ecosystem – they will complement rather than replace.
Functional Comparison: Keyword Search vs. AI Models
Keyword search has proven effective for retrieving documents and has continuously evolved since the internet's emergence in the 1990s. Crucially, this method enables users to discern the document's origin and the creator's identity.
However, AI models aren't designed to do that yet. They make predictions based on input, but their output is an abstraction of source material. While efforts are underway to address these limitations, numerous situations exist where authorship, trust, and accountability are paramount, and an uncited ChatGPT summary won't suffice.
Another crucial aspect is that search engines rely heavily on advertising revenue generated from visitor traffic. When an AI model delivers direct answers within a specific company's interface, without directing users to other websites, monetization options differ, and profits may not be evenly distributed.
AI development is expensive and will likely become more so as data owners tighten control to prevent unauthorized training and start hammering out licensing agreements. Currently, OpenAI, Google, Microsoft, and others rely on subscription revenue and partnership deals. However, the number of customers has limitations. The solution is an inclusive system like web advertising, despite its flaws.
If AI search is to challenge web search, it needs a business model that doesn't just rehash others' content without compensation, like providing summaries, code snippets, or visual remixes, and selling the results at a markup.
OpenAI and its partners can't expect every digital content creator to relinquish their work to be used in AI models that benefit the model developers but offer no upside for other stakeholders. Those whose work has been integrated into existing AI models also have a right to be heard, if the law permits.
Ironically, OpenAI reportedly sent copyright infringement takedown demands to the ChatGPT subreddit for using the company's logo. However, the subreddit moderators clarified OpenAI had permitted use of the branding. Hopefully, this dispute will be amicably resolved.
0 comments:
Post a Comment