Monitoring Desk
ISLAMABAD/MOUNTAIN VIEW: As Alphabet Inc (GOOGL.O) looks past a chatbot blunder that cost it $100 billion in market value, another issue has emerged from its efforts to incorporate generative artificial intelligence into its popular Google Search: the cost.
Executives in the technology sector are debating how to operate AI such as ChatGPT while accounting for the high costs. The wildly popular OpenAI chatbot, which can draught prose and answer search queries, has “eye-watering” computing costs of a couple or more cents per conversation, according to the startup’s CEO Sam Altman on Twitter.
Alphabet Chairman John Hennessy told Reuters in an interview that an exchange with AI known as a large language model will likely cost ten times more than a standard keyword search, though fine-tuning will help reduce the cost quickly. Even if revenue from potential chat-based search ads is generated, analysts believe the technology could cost Alphabet several billion dollars in additional costs. In 2022, its net income was nearly $60 billion.
Morgan Stanley estimated that Google’s 3.3 trillion search queries cost about a fifth of a cent each last year, a figure that would rise depending on how much text AI must generate. Analysts estimate that if ChatGPT-like AI handles half of the queries it receives with 50-word answers, Google’s expenses could rise by $6 billion by 2024. Google is unlikely to require the use of a chatbot to handle navigational searches for sites such as Wikipedia. Others arrived at a similar bill in a variety of ways. For example, SemiAnalysis, a chip technology research and consulting firm, estimated that adding ChatGPT-style AI to search could cost Alphabet $3 billion, a figure limited by Google’s in-house chips known as Tensor Processing Units, or TPUs, as well as other optimizations.
Artificial intelligence more expensive than traditional search
The computing power required makes this type of AI more expensive than traditional search. According to analysts, such AI requires billions of dollars in chips, which must be spread out over their useful life of several years. Electricity also adds costs and pressure to businesses that want to reduce their carbon footprint. The process of handling AI-powered search queries is known as “inference,” in which a “neural network” loosely modelled on the biology of the human brain infers an answer from prior training.
In contrast to a traditional search, Google’s web crawlers have scanned the internet to compile an index of information. When a user enters a query, Google returns the most relevant answers from the index. Alphabet is under pressure to accept the challenge despite the cost. Similarweb estimates that Microsoft Corp (MSFT.O) held a high-profile event earlier this month at its Redmond, Washington headquarters to show off plans to embed AI chat technology into its Bing search engine, with top executives taking aim at Google’s search market share of 91%.
A day later, Alphabet talked about plans to improve its search engine, but a promotional video for its AI chatbot Bard showed the system answering a question inaccurately, fomenting a stock slide that shaved $100 billion off its market value. Microsoft later drew scrutiny of its own when its AI reportedly made threats or professed love to test users, prompting the company to limit long chat sessions it said “provoked” unintended answers.
Amy Hood, Microsoft’s Chief Financial Officer, told analysts that the benefits of gaining users and advertising revenue outweighed the costs of launching the new Bing to millions of users. Another Google competitor, Richard Socher, CEO of search engine You.com, stated that adding an AI chat experience as well as applications for charts, videos, and other generative tech increased costs by 30% to 50%.
A Google source cautioned that it is too early to estimate how much chatbots might cost because efficiency and usage vary greatly depending on the technology involved, and AI already powers products such as search. Nonetheless, footing the bill is one of two main reasons why search and social media behemoths with billions of users have not rolled out an AI chatbot overnight, according to Accenture’s chief technology officer, Paul Daugherty.
