Key Takeaways
- In China, a man was arrested for allegedly using the AI chatbot ChatGPT to spread fake news about a train crash, raising concerns about the political implications of AI-generated content.
- The man used ChatGPT to generate multiple versions of the fake news article, which he posted on the Chinese blogging platform Baijiahao, bypassing content restrictions.
- This arrest may be one of the first cases under China's new anti-AI guidelines, which aim to restrict the use of AI for creating deepfakes and limit AI-generated news to approved outlets.
- OpenAI, the company behind ChatGPT, employs underpaid data labelers who play a crucial role in training machine learning models but receive low wages and no benefits. They are paid as little as $15 per hour.
- Previously, OpenAI outsourced moderation work to Africa, where workers were paid as low as $2 per hour. This practice faced criticism and legal action.
- These incidents highlight the ethical challenges of AI, including the potential for misinformation and the need to ensure fair treatment and compensation for the human workforce involved in AI development.
- Responsible and sustainable advancement of AI technology requires addressing these challenges and protecting the rights of individuals contributing to AI progress.
As AI technology continues to advance and impact our daily lives, it is also becoming a tool for spreading misinformation and a source of labor concerns.
Recently, a man in China was arrested for allegedly using the AI chatbot ChatGPT to generate and spread fake news about a train crash, raising questions about the political implications of AI-generated content.
Meanwhile, a report by NBC News reveals the underpaid and undervalued workforce behind AI technology, as data labelers are paid a mere $15 per hour for their crucial role in training machine learning models.
Man Arrested in China for Alleged Misinformation Campaign Using ChatGPT
Chinese police recently arrested a man in Gansu province for allegedly using the AI chatbot ChatGPT to spread misinformation.
The accused, identified by his surname Hong, reportedly created fake articles about a train crash that killed nine people on April 25.
He is charged with “picking quarrels and provoking trouble,” a political catch-all offense often used against dissidents and activists, which can result in a five-to-10-year prison term.
This arrest could be one of the first cases under China’s new anti-AI guidelines, which restrict people from using AI to create deepfakes and limit news stories created by AI to a list of approved outlets.
Critics argue that repressive governments might use the “it’s all just AI” excuse to discredit legitimate reports.
The Role of ChatGPT in Content Generation
Hong is accused of using ChatGPT to generate multiple versions of the fake news article, which he then posted on Chinese blogging platform Baijiahao.
He allegedly bypassed Baijiahao’s restrictions on posting the same content to multiple accounts, garnering around 15,000 clicks.
Hong’s company, which operates multiple blog-type outlets, is registered in Shenzhen, a major tech business and manufacturing hub in southern China.
Chinese users can access ChatGPT through a VPN, but local companies like Baidu are developing their own chatbots that comply with the country’s stringent censorship policies.
Baidu also owns the Baijiahao blogging platform.
Hong is accused of using ChatGPT to generate multiple versions of the fake news article, which he then posted on Chinese blogging platform Baijiahao.
Invisible Workforce Behind AI: Underpaid Data Labelers
A report by NBC News reveals that OpenAI, the company behind ChatGPT, pays a large number of US contractors as little as $15 per hour to perform data labeling.
This crucial task involves parsing data samples to help automated systems better identify items within the dataset.
Data labelers tag specific items so that machines can learn to recognize them independently. Despite its importance in training machine learning models, data labelers receive no benefits and are paid little more than minimum wage in some states.
One data labeler, Alexej Savreux, states, “You can design all the neural networks you want, you can get all the researchers involved you want, but without labelers, you have no ChatGPT. You have nothing.”
Previous Outsourcing to African Moderators
Before employing US contractors, OpenAI outsourced its moderation work to Africa, where it could pay workers as low as $2 per hour due to lower wages and limited labor laws.
The company previously collaborated with Sama, an American firm that connects big tech companies with low-wage contractors in Third World countries.
Sama has faced a lawsuit and accusations of providing poor working conditions.
These low-paid moderators helped OpenAI build a filtration system for ChatGPT, which screens out offensive material.
To achieve this, the moderators had to endure exposure to disturbing content, including descriptions of murder, torture, sexual violence, and incest.
The recent incidents in China and the United States highlight the complex issues surrounding AI technology, from its potential misuse for spreading misinformation to the labor concerns of the human workforce behind it.
As AI continues to permeate various aspects of society, it is crucial to consider the ethical implications of its use and development, as well as the need to protect the rights of the individuals who contribute to its progress.
Addressing these challenges will be essential for ensuring the responsible and sustainable advancement of AI technology in the coming years.
