Chinese geo chatbot accused of censorship, bias

Occurred: June 2024

GeoGPT, a new AI chatbot backed by the International Union of Geological Sciences (IUGS), has been criticised by geologists over potential censorship and bias and poor transparency.

Developed under the Deep-time Digital Earth (DDE) programme, which receives substantial funding from China, GeoGPT is meant to assist geoscientists, particularly in developing countries, by providing access to extensive geological data

However critics, including Professor Paul Cleverley, have pointed out potential censorship and lack of transparency in GeoGPT’s responses. While DDE representatives claim that the chatbot’s information is purely geoscientific and free from state influence, tests with Qwen (developed by Alibaba), its underlying AI model, suggest that certain sensitive questions may be avoided or answered inadequately. 

For instance, when asked about fatalities in a Ghanaian mining operation run by the Shaanxi Mining Company, Qwen cited outdated information and avoided specifics, whereas ChatGPT provided detailed information on the incident.

Concerns were also expressed about the bot's Chinese funding and the potential for biased data usage. Geoscientific research, including valuable information about natural resources, could be strategically filtered. 

The terms of use for GeoGPT prohibit generating content that undermines national security or incites subversion, aligning with Chinese laws, which may influence the chatbot’s outputs. 

IUGS president, John Ludden, said GeoGPT’s database will be made public once appropriate governance is ensured, but geoscientists remain skeptical about the impartiality and transparency of GeoGPT’s data and responses.

Operator: 
Developer: Deep-time Digital Earth
Country: China
Sector: Energy
Purpose: Support geological research
Technology: Chatbot
Issue: Bias/discrimination; Copyright; Human/civil rights; Transparency