Google’s AI fight is moving to new ground - FT中文网
登录×
电子邮件/用户名
密码
记住我
请输入邮箱和密码进行绑定操作:
请输入手机号码,通过短信验证(目前仅支持中国大陆地区的手机号):
请您阅读我们的用户注册协议隐私权保护政策,点击下方按钮即视为您接受。
谷歌

Google’s AI fight is moving to new ground

The company has defended search but its dominant position in online advertising may still be up for grabs

To judge from its latest quarterly earnings, Google has shored up its defences against the first wave of disruption from generative artificial intelligence. But the real challenge has barely begun.

The search giant misfired with its first, panicked response in the wake of ChatGPT (if you never used Bard, you aren’t alone), but this year it seems to have found its feet. People seem to like having AI-generated answers spliced into their search results. According to Google, they have responded by thinking up 10 per cent more of the types of queries that produce these summaries.

Being presented with ready-made answers means they are less likely to click on links, of course — according to Pew Research in the US, about half as likely. But that hasn’t stopped solid growth in search advertising revenue.

Defending search has been the first order of business. But generative AI promises to draw people’s time and attention with a whole new set of tools, of which chatbots such as OpenAI’s ChatGPT are only the first. Google’s massive advertising business, which accounts for about a third of the entire global digital ads market, was built on indexing and guiding people around the web. Much of this may soon be up for grabs.

For many people, ChatGPT has already broken the habit of turning to a search engine as the default action every time they need something online. AI-powered apps with a wider set of capabilities are becoming the centre of the action.

Google chief executive Sundar Pichai seemed to confirm this view in a call with Wall Street analysts this week. AI in search was “more information-focused”, he said, while his company’s separate AI app was “more your assistant, [a] more personal, proactive and powerful assistant for every aspect of your daily life”.

This points to a bifurcation not just in the things that people want to do, but also where they take those actions, leading them to turn from a browser-based search engine towards standalone apps. For Google, browsers have been an important stronghold, and placement as the default search service in browsers is a cornerstone of its business. With AI apps, the field is wide open.

Google reports respectable numbers for its entry in this category: about 450mn people visit its Gemini app at least once a month, roughly half as many as those who go to ChatGPT. 

But on the metrics that really matter, OpenAI has a strong lead. According to estimates that Google itself produced during its antitrust trial in the US earlier this year, ChatGPT had nearly five times as many daily users, each of them making, on average, twice as much use of the app as users of Gemini.

There is no guarantee OpenAI’s impressive grip on the popular imagination or its momentum will last. After some early slips, Google’s AI models rank just as highly on many tests. It seems a pure technology edge here will be hard for anyone to maintain for long.

The search company’s extensive collection of platforms and products — it boasts of 15 services with at least 500mn users each — also gives it a powerful weapon to draw users to its AI, and to generate the data to refine its service.

But this is only part of the challenge. Google also needs to find different ways to make money.

A turn towards charging subscriptions for premium search services and surging demand for its cloud computing platform are two bright spots. If AI agents guide and shape more of what people do online, reducing their propensity to click, search advertising will atrophy. Rather than monetising attention, Google will need to monetise actions.

To judge by its latest moves, OpenAI has also decided this is where the money will be. Last week saw the launch of ChatGPT agent, a service designed to click around the web on behalf of a user. As reported by the Financial Times, OpenAI is also getting ready to take a cut from commercial transactions that take place over ChatGPT.

The potential to combine these tools — using an agent to steer people towards, or even make, their next purchase, then taking a commission — points to how OpenAI could become a power on the commercial web.

Google’s investors can at least draw heart from signs that their company is starting to find its innovative spark. Project Mariner, a prototype it showed off two months ago, closely echoes ChatGPT agent.

But the lock on advertising that Google has long enjoyed thanks to search is starting to loosen, leaving it to fight on a new battlefield against AI apps — and not just those from OpenAI.

richard.waters@ft.com

版权声明:本文版权归FT中文网所有,未经允许任何单位或个人不得转载,复制或以任何其他方式使用本文全部或部分,侵权必究。

谷歌为Anthropic打造的2000亿美元华尔街融资机器

私募信贷、芯片租赁和数据中心担保,支撑起AI支出的全新庞大模式。

“日元干预”等于“美国自保”

美国联手日本支撑日元,不只是出于盟友情谊,更是为了避免日本加息或美国国债遭抛售、导致美债收益率进一步走高。

“诅咒之岛”:科技游民与诈骗犯藏身的千亿美元奢华开发项目

警方的突击搜查再次打击了马来西亚陷入困境的中资“森林城市”项目的声誉。

问题不在因凡蒂诺

马杜罗:应该将国际足联的监管职能与商业活动分开,其治理应真正做到包容并具有代表性,监督必须真正独立。

俄罗斯扩大“影子”液化天然气船队,应对欧盟禁令

随着明年制裁进一步收紧,越来越多的“影子”船舶将帮助俄罗斯继续出口液化天然气。

阿斯利康与百时美施贵宝:大药企有时也不够大

当资产负债表规模扩大、能够押注潜在重磅药物时,规模才会带来优势。
设置字号×
最小
较小
默认
较大
最大
分享×