Baidu shares fell as much as 10 percent on Thursday after the web search company showed only a pre-recorded video of its AI chatbot Ernie in the first public release of China’s answer to ChatGPT.
The Beijing-based tech company claimed that Ernie would rebuild its business and talked about plans to incorporate generative artificial intelligence into its search engine and other products.
But on Thursday, millions of people tuning into the event were left with little idea whether Baidu’s chatbot could compete with ChatGPT.
During a highly publicized and eagerly anticipated news conference for Ernie, Baidu founder Robin Li stood next to an open chat screen, describing questions previously typed into the chatbot. He admitted that the company was only showing a demo of the technology that it had previously developed.
Lee said some users will soon be able to test Ernie on their own, but did not provide a timeline for a full public rollout. The company is starting with a limited public release to business partners.
Ernie’s planned launch comes as US groups such as OpenAI and Google continue to advance their generative AI development. OpenAI this week introduced GPT-4, its latest AI model that it claims can beat some humans on tough professional tests like the US bar exam.
Microsoft is incorporating GPT-4 into Bing and other products, while US corporations are testing the technology in their workplaces and products.
With nearly every Chinese tech giant racing to introduce a homegrown version of ChatGPT, analysts say they believe Baidu’s years of investment in AI and development of natural language learning technology have given the group an early lead.
“Sometimes when we use it, we’re pleasantly surprised, sometimes we think there’s an obvious mistake,” Lee told the audience. “But one thing’s for sure, it’s moving very quickly.”
However, there was disappointment in the live comments posted by many of the two million people watching Li’s speech on the WeChat social media app. It didn’t take long for Baidu’s share price decline to dominate the minute-by-minute quips feed.
“Its extremely strong ability to understand and express language will allow any company to get closer to their customers,” Lee said. “It’s an opportunity for every company, and it’s going to have an impact on everyone, too.”
At the end of the roughly 45-minute performance, a chorus of WeChat users chimed in with variations of: “Enough?” Two Baidu employees said they were equally disappointed with the rollout.
For months, the company diverted internal resources to building Ernie, the two people said, including reallocating most of Baidu’s servers running Nvidia’s powerful A100 chips to the AI team.
OpenAI’s introduction of ChatGPT last November caught Baidu off guard, two people close to the company’s AI efforts said, adding that they didn’t believe the US group had significantly superior technology until then.
Its launch set Baidu scrambling to catch up and undertake the grueling process of fine-tuning Ernie with human feedback, a critical step in improving ChatGPT’s accuracy and fluidity, people close to Baidu said.
The training program requires humans to grade thousands of responses spat out by AI models, helping them gradually score higher and improve. One employee said the relative lack of OpenAI meant Baidu’s team had to experiment on its own.
The US group also said it would not disclose details on the technical underpinnings of the GPT-4, which is a change from previous models such as the GPT-3, on which it had released some information.
“We can only explore by ourselves. It took OpenAI more than a year to train ChatGPT, and another year to tune GPT-4,” said a Baidu employee. “That means we’re two years behind.”
Baidu did not immediately respond to a request for comment.
Additional reporting by Nian Liu in Beijing