The intelligence is synthetic. The bullishness it engenders is natural. Generative AI chatbots may allow humans to interact seamlessly with computers. OpenAI’s popular ChatGPT programme has therefore inspired optimism in tech, dispelling gloom around defunct Silicon Valley Bank.
Google just launched its Bard chatbot in rivalry with the upgraded ChatGPT-4 AI unveiled by OpenAI last week. For the moment, ChatGPT-4 enjoys the advantage of coming to market first.
Over 100mn people have signed up to use ChatGPT, according to estimates. Many work in tech, which is why so many experiments centre on generating code for video games.
Companies are starting to integrate the tools. Backer Microsoft is going to add ChatGPT-4 to paid-for Office apps like Excel spreadsheets. LinkedIn will use it to power prompts so users interact with articles. Any work that involves text or images, from software engineers to lawyers to HR, can make use of generative AI.
Consumerisation is moving at a rapid pace. Monetisation is still a work in progress. Costs are high and highly variable. Revenue is low.
OpenAI offers a $20 per month subscription for ChatGPT Plus and charges developers to license its tech. It is reported to forecast $200mn in revenue this year. If it secures a $29bn valuation, it would be valued at 145 times sales.
Big Tech could grab some of OpenAI’s revenue by including AI in free products. Google is integrating generative AI into tools like word processor Google Docs. Meta is using it to improve its advertising system.
AI is only as good as the information in its large language models. Creating these models could be one fruitful source of revenue. AI firm Anthropic’s latest model Claude is already in use by customers.
OpenAI is clear that the tech is still flawed. Microsoft’s AI-powered search engine has produced some odd results. Companies that hastily add imperfect AI features risk turning users against them. But the consequences of moving too slowly will be worse.