The US is at risk of losing its edge to China in the race for artificial intelligence (AI), according to Alexandr Wang, the CEO of Scale AI. Wang was speaking at a summit organized by the company for government officials, Bloomberg reported.
AI is among the future technologies that countries around the world are looking to establish their supremacy in as the world evolves into the next age. Interesting Engineering has previously reported that China has been denting U.S. dominance in the technology sector. In studies conducted by policy institutes, China has already established monopolies in certain areas.
Not only is China home to the leading research institutes of a large spectrum of topics, but it is also generating nearly half of the world’s most impactful papers in these areas. When it comes to AI, the U.S. is still ahead but at risk of losing its lead, according to Wang.
What does Scale AI do?
The U.S. lead in AI has been demonstrated by the unveiling of ChatGPT, the conversational chatbot that multiple companies in China have tried to replicate but without the resounding popularity.
While OpenAI is credited with creating GPT, the large language model (LLM) behind ChatGPT, the role of companies like Scale AI cannot be discounted. LLMs need tons of data to train the bots, and the bot is only as good as the data it is trained on.
In 2016, Alexander Wang dropped out of MIT and founded Scale AI to help companies get the right training data to train their models. Apart from OpenAI, the company is also helping chipmaker Nvidia, carmaker Toyota, and the U.S. government build better AI systems.
How the U.S. is losing its edge
According to Wang, AI is unavoidable technology that must be integrated into military operations to stay ahead of adversaries. Comparing AI to nuclear weapons, Wang said that the technology would reshape global diplomacy and power.
Wang also pointed out that China was investing heavily in AI, both in terms of absolute numbers as well as relative to its defense budget. In a presentation at the summit, Wang detailed that the People’s Liberation Army had invested $1.6 billion in the technology in 2020, compared to the $1.3 billion allocated by the Department of Defense.
Senator Mike Rounds, speaking at the same event, agreed that the U.S. currently held the advantage when it came to data collection and labeling but also warned that services like TikTok, which is owned by Chinese company ByteDance could give the Chinese access to additional English language samples to train their AI systems.
Interesting Engineering has also reported that the flurry of AI models developed by Chinese tech counterparts after ChatGPT’s success has also been bilingual and intended for users outside China.
Wang further added that the U.S. must leverage the vast amounts of data it collects through its military hardware, such as sensors, cameras, and satellites, to train AI. Ensuring the quality of the inputs can turn the “hardware advantage into data advantage.”
