According to Baidu CEO Robin Li, there will be a big disruption in the AI sector and very few businesses would be able to weather what he perceives as an imminent “AI bubble.” Li compared the current wave of AI businesses to the dot-com bubble of the 1990s when speaking at the Harvard Business Review Future of Business Conference. She predicted that when the bubble busts, 99 percent of AI ventures will fail. Only one percent of businesses, he pointed out, will endure and go on to produce significant value for society.
Li highlighted how large language models (LLMs) have advanced significantly in the last 18 to 20 months, especially in terms of the precision of the responses they generate. He noted that while early iterations of these models frequently experienced hallucinations, resulting in inaccurate or misleading information, this problem has been much mitigated in more recent iterations. Li stated, “That problem has been largely solved,” suggesting that consumers may now more safely trust the responses given by frontier AI models.
The CEO of Baidu cautioned that most of the hype surrounding AI might not last long, despite major advancements in the field. He clarified that although a small number of exceptional businesses would eventually prosper, the bulk are trapped in a “inevitable bubble” and will find it difficult to endure over the long haul. Li predicted that it would take a further 10 to 30 years before AI starts to largely displace human labor in a variety of industries.
Li advised governments, organizations, and people to get ready for the profound changes that will occur with AI technology as it continues to progress. He emphasized that, as AI continues to transform industries and daily life, it is critical to adjust to the impending changes in the job market and the economy as a whole. Although he is still confident in AI’s ability to advance society, his forecast highlights the instability and significant risks that the AI sector is currently confronting.