Chinese AI developer claims it’ll beat ChatGPT in a few months

Joel Loynds
SparkDesk AI announcement

The AI developer, iFlytek, took the stage claiming that they’ll be able to best OpenAI in Chinese and hit English standards for ChatGPT in just a few short months.

iFlytek, a Chinese AI developer which specializes in language tools, has begun to launch its own AI chatbot similar to ChatGPT and Google Bard. Known mainly for its development of translation and voice recognition software, iFlytek’s latest product is called “SparkDesk”.

Essentially the same style of AI chatbot as its rivals, iFlytek doesn’t intend on sticking around as a Chinese alternative. In the presentation, the chairman of the company, Liu Qingfeng, claimed that they’ll be able to surpass OpenAI’s Chinese efforts by October. The company is also aiming to meet the English standards that ChatGPT currently has in this period of time too.

SparkDesk is not the first Chinese AI bot to come out, with Ernie Bot currently sitting at the top of the pile. However, Ernie Bot’s initial reception was poor, as the developers, Baidu, only demonstrated it with pre-recorded videos.

iFlytek performed a live demo and even allowed users in the audience to ask it questions via a voice prompt. As of right now though, iFlytek and other Chinese AI developers have a different hurdle to tackle.

US and Chinese restrictions on AI development

SparkDesk AI

Despite the lofty claims from iFlytek that its AI will be able to surpass ChatGPT in Chinese, the company is still under strict sanctions from 2019. This prevents them from buying US parts to help with its development.

SparkDesk, Ernie Bot, and others are also regulated a little more heavily than over in the West. While Nvidia and certain government bodies are attempting to put up guardrails or investigate privacy concerns, China has already drafted regulations around generative AI.

Key components of the legislation include that the data must be accurate, cannot insight hatred, and cannot be against the state’s imposed restrictions:

“The use of generative artificial intelligence generated content should reflect the core values of socialism, shall not contain subversion of state power, overthrow the socialist system, incitement to split the country, undermine national unity, promote terrorism, extremism, promote ethnic hatred, ethnic discrimination, violence, obscene pornography, false information, and may disrupt the economic order and social order of the content.”

Source, translated by DeepL
Sign up to Dexerto for free and receive:
Fewer Ads|Dark Mode|Deals in Gaming, TV and Movies, and Tech