The Cyberspace Administration of China has new policies on the creation of AI tools.
China’s regulators are looking into Generative AI tools. They want to ensure tech companies are managing the quick-developing Generative AI tools. The move came after trailblazer Alibaba revealed plans to release its own chatbot.
Generative AI is one of the most common types of artificial intelligence technology. These AI tools can create different types of content. It can include audio, imagery, and text. This is always done by analyzing large quantities of data.
One of the most popular generative AI is ChatGPT. It was then rolled out in November 2022 by OpenAI, the American AI lab. Knowledge of ChatGPT’s capabilities spread like wildfire. The tool was soon known for responding to questions using human-like language.
ChatGPT also showed it had a wide array of abilities. It could summarize complex research studies in layman’s terms. It can answer logical questions and generate essays. The chatbot also even solved medical exams that were vital for students. This has led to other companies coming out with their own versions of ChatGPT.
Alibaba Throws Hat in the AI Ring
Alibaba revealed earlier this week its plans to join in the generative AI race. The Chinese tech giant announced it’s working on a chatbot to rival ChatGPT. The company is calling it Tongyi Qianwen.
There's a rough translation of what Tongyi Qianwen means. It means “to look for an answer by asking thousands of questions.” There’s no news if Alibaba will even have an English name for their bot.
Alibaba said the chatbot's set to become integrated into its system. It didn’t mention any specific timeline but said it’s happening soon.
Tongyi Qianwen will reportedly work in both Chinese and English. The initial plan is to integrate it into DingTalk. This is the company’s workplace messaging app. The bot will be able to do various tasks. For example, it can turn meeting conversations into written notes. It can also draft business proposals and write emails.
Alibaba also plans to integrate it into Tmall Genie. It’s an app akin to Alexa, Amazon’s voice assistant smart speaker.
The company said integrating its ChatGPT rival model will improve the user’s experience. Alibaba developers and customers will receive access to their AI model. It will help them create customized AI features in an affordable manner.
Alibaba Cloud will also give its clients access to Tongyi Qianwen. This will help them develop customized and large-scale language models. The chatbot will become calibrated with premier proprietary information and data from clients. This will also help lower costs and save on resources. Alibaba Cloud’s enterprise customers in the country can beta test Tongyi Qianwen.
Alibaba Group chairman Daniel Zhang said the company is at a watershed moment. He said the moment is now driven by generative AI and cloud computing. He admitted that businesses across all sectors are already beginning to embrace AI. Especially if they want to remain in the game.
Alibaba isn’t the only company dipping its toes in the generative AI pool. The Chinese company Baidu and Google have also said they’re working on their own version of AI chatbots.
China Drafts New AI Regulations
Alibaba’s decision to come up with its own version of ChatGPT is not surprising. So is the Chinese government’s latest move. It will introduce regulations to manage this type of technology in the country.
The recent draft measures lay down the groundwork for companies developing AI services. The policies will guide companies on the kind of content they can generate. They also have to make sure the data used to teach the AI models are not discriminating. The models must not discriminate against people based on their race and gender. It should not work against certain ethnicities as well.
The CAA says content created by future AI tools must reflect China’s core values. It must not encourage subversion of the government’s power.
China new AI rules also say companies are responsible for the validity of their data. These are all used to train the algorithm. Regulators also underlined the importance of AI chatbots not producing false information.
The proposed law is already slated to roll out anytime this year. The regulations come as many China-based tech companies unveiled new generative AI products.
The CAC also stated service providers must have users provide their real names. Providers that fail to follow the regulations will face a fine. They will also have their services suspended. Their failure to comply could even lead to a criminal investigation.
The regulator also revealed measures to deal with inappropriate content. It said companies must update their technology within three months after the incident. This is to stop the creation of similar content.
Some AI specialists have issued warnings on chatbots. They said tools like ChatGPT can give credible-sounding responses. But the replies are incorrect and the mistakes in some questions are glaring. They noted chatbots might even be confident about the incorrect answers they provided. These will mislead users and can even result in widespread copyright infringement.
The wonders of generative AI haven’t impressed everyone though. Some universities in Japan have limited the students’ use of ChatGPT. Sophia University in Tokyo is one. The school has limited the use of the chatbot due to concerns about data leaks from using this tool.
Italy has already banned ChatGPT. It’s the first Western country to do so. The country is also set to investigate the AI service over privacy issues. Germany is reportedly also thinking of banning ChatGPT over similar concerns.