How Siri, Alexa and Google Assistant lost out in the A.I. race

On a rainy Tuesday in San Francisco, Apple executives took the stage in a crowded hall to unveil the fifth-generation iPhone. The phone, which looked identical to the previous version, had a new feature that soon got the whole audience talking: Siri, a virtual assistant.

Scott Forstall, then Apple's head of software, pressed the iPhone button to summon Siri and began asking her questions. At his request, Siri checked the time in Paris ("8:16 p.m.," Siri replied), defined the word "mitosis" ("Cell division in which the nucleus divides into nuclei containing the same number of chromosomes," she said), and pulled up a list of 14 highly rated Greek restaurants, five of which were in Palo Alto, California.

"I've been in the A.I. field for a long time, and it still amazes me," Mr. Forstall said.

That was 12 years ago. Since then, Siri and competing artificial intelligence assistants like Amazon Alexa and Google Assistant have fallen far short of wowing people. The technology has largely remained unchanged, and talking assistants have been the subject of jokes, including a 2018 "Saturday Night Live" sketch that featured a smart speaker for seniors.

The tech world is now marveling at a different kind of virtual assistant - chatbots. These artificial intelligence-enabled bots, such as ChatGPT and the new ChatGPT Plus from San Francisco-based OpenAI, can quickly improvise answers to questions typed in a chat room. People have used ChatGPT for complex tasks such as coding software, writing business proposals, and writing fiction.

And ChatGPT, which uses artificial intelligence to guess which word is next, is improving rapidly. A few months ago, it couldn't write a proper haiku; now it can do so with enthusiasm. On Tuesday, OpenAI unveiled its next-generation A.I. engine, GPT-4, on which ChatGPT runs.

The hype surrounding chatbots illustrates how Siri, Alexa and other voice assistants that once generated similar enthusiasm have squandered their edge in the A.I. race.

Products have faced obstacles over the past decade. Siri has faced technological hurdles, including clunky code that took weeks to update basic features, said John Burkey, a former Apple engineer who worked on the assistant. Amazon and Google miscalculated how voice assistants would be used, leading them to invest in areas where the technology rarely paid off, former employees said. When those experiments failed, the companies' enthusiasm for the technology waned, they said.

Voice assistants are "dumb as a rock," Satya Nadella, Microsoft's chief executive, told the Financial Times in an interview this month, saying that newer A.I.'s will lead the way. Microsoft has worked closely with OpenAI, investing $13 billion in the startup and incorporating its technology into the Bing search engine as well as other products.

Editor's Choice

36 hours in Taipei, Taiwan

How do I get my daughter to accept my boyfriend after eight years?

Is it safe to treat my neck with a chiropractor?
Apple declined to comment on Siri. Google said it intends to create a great virtual assistant to help people on their phones, in their homes and in their cars; the company is separately testing a chatbot called Bard. Amazon has said that the number of customers using Alexa has increased by 30 percent worldwide over the past year, and that it is optimistic about its mission to create a world-class A.I.

Assistants and chatbots are based on different kinds of A.I. Chatbots work on so-called big language models, which are systems trained to recognize and generate text based on huge datasets taken from the Internet. They can then suggest words to complete a sentence.

In contrast, Siri, Alexa and Google Assistant are essentially so-called command and control systems. They can understand a limited list of questions and requests, such as "What's the weather in New York?" or "Turn on the bedroom light." If a user asks a virtual assistant to do something that isn't in their code, the bot simply says it can't help.

Siri also had a cumbersome design that made it time-consuming to add new features, said Mr. Burki, who in 2014 was tasked with improving Siri. Siri's database contains a gigantic list of words, including names of music artists and places, such as restaurants, in nearly two dozen languages.

That turned it into "one big snowball," he said. If someone wants to add a word to the Siri database, he added, "it goes into one big pile."

So seemingly simple updates, such as adding a few new phrases to a dataset, would require rebuilding the entire database, which could take up to six weeks, Mr. Burkey said. Adding more complex features, such as new search tools, could take almost a year. That means Siri has no path to becoming a creative assistant like ChatGPT, he said.

According to former Amazon and Google executives, Alexa and Google Assistant use technology similar to Siri, but the companies have failed to generate significant revenue from these assistants. (In contrast, Apple has successfully used Siri to attract customers to its iPhones.)

After Amazon launched Echo, an Alexa-powered smart speaker, in 2014, the company hoped the product would help it boost sales at its online store by allowing customers to talk to Alexa to place orders, said a former Amazon executive with ties to Alexa. But while people have enjoyed playing around with Alexa's ability to respond to weather prompts and set alarm clocks, few have asked Alexa to order merchandise, he added.

Amazon may have invested too much in producing new types of equipment, such as now-discontinued alarm clocks and microwave ovens that work with Alexa, which sold at or below cost, the former executive said.

The company has also underinvested in creating an ecosystem that makes it easy for people to expand Alexa's capabilities, similar to what Apple did with its App Store, which fueled interest in the iPhone, this person said. While Amazon offered a "skills" store to make Alexa control third-party accessories like light switches, it was difficult for people to find and customize skills for speakers - unlike the easy experience of downloading mobile apps from app stores.

"We never had that App Store moment for assistants," said Carolina Milanesi, a consumer technology analyst at Creative Strategies, a research firm that has been a consultant to Amazon.

Late last year, Amazon's Alexa division was the main target of 18,000 employee layoffs, and a number of top Alexa executives left the company.

Kinley Pearsall, an Amazon spokeswoman, said Alexa is more than a voice assistant and "we are just as optimistic about this mission."

Amazon's misfires with Alexa may have derailed Google, says a former manager who worked on Google Assistant. Google engineers spent years experimenting with its assistant to mimic Alexa's capabilities, including developing smart speakers and voice-controlled tablet screens to control home accessories such as thermostats and light switches. The company later integrated advertising into these home products, which did not become a major source of revenue.

Over time, Google realized that most people only use the voice assistant for a limited number of simple tasks, such as starting timers and playing music, the former manager said. In 2020, when Prabhakar Raghavan, one of Google's top executives, took over Google Assistant, his group refocused the virtual assistant as a key feature for Android smartphones.

In January, when Google's parent company cut 12,000 employees, the team working on operating systems for home devices lost 16 percent of its engineers.

Many big tech companies are now scrambling to come up with an answer to ChatGPT in a hurry. Last month, Apple's headquarters hosted its annual A.I. Summit, an internal event for employees to learn about its big language model and other A.I. tools, said two people briefed on the program. Many engineers, including members of the Siri development team, tested language generation concepts each week, the people said.

Google also said Tuesday that it will soon release generative A.I. tools to help businesses, governments and software developers build applications with embedded chatbots and incorporate basic technology into their systems.

In the future, chatbot and voice assistant technologies will merge, A.I. experts say. That means people will be able to control chatbots through speech, and users of Apple, Amazon and Google products will be able to ask virtual assistants to help them with their work, not just to check the weather.

"Before, these products didn't work because we didn't have the ability to have human-level dialog," says Aravind Srinivas, founder of Perplexity, an A.I. development company that offers a chatbot-based search engine. "Now we have it."