Artificial intelligence and new policies

Artificial intelligence and new policies

The use of ChatGPT, a conversational robot powered by OpenAI, has spread to academic and professional circles, with politicians and legislators using it to create speeches and laws. Some governments are even attempting to use the robot for their own purposes, while others worry that it could be weaponized for influence campaigns.

Examples of ChatGPT's use include a Japanese parliamentarian who used it to question the prime minister in March, and the robot drafting an amendment to the 2024 Olympic Games bill in France. French President Emmanuel Macron even referenced ChatGPT on Twitter, praising its ability to consider Europe competitive in the innovation race.

However, the American technology behind ChatGPT was not designed to make judgments, but rather to respond with the most appropriate words to a request, which could lead to it taking opposite positions on the same issue.

Despite its popularity with more than 100 million active users just two months after its launch earlier this year, ChatGPT has also faced criticism. The general secretary of the CGT union, Sophie Binet, ridiculed Macron for using the robot, suggesting that the president's statements to defuse the social crisis caused by pension reform "could have been made by ChatGPT."

Pascal Marchand, a professor of information sciences at the University of Toulouse, suggests that AIs like ChatGPT are capable of generating speeches that are faithful to traditional ideological markers of politicians. However, due to their inability to innovate, they may be less relevant for parties seeking to adapt their discourse to changing times.

Right-wing parties believe that ChatGPT is steeped in the liberal and progressive values of Silicon Valley, while the president of France's National Rally party has raised concerns about an "artificial intelligence great replacement" on networks, referring to an alleged demographic replacement plan attributed to waves of migration to Europe by some far-right sectors.

OpenAI and its competitors have biases due to training from a large corpus of text and filters to limit objectionable comments. In New Zealand, researcher David Rozado designed the RightWingGPT robot, an AI trained to produce conservative arguments.

Elon Musk, head of Twitter and an investor in OpenAI, has expressed interest in launching TruthGPT, a less "politically correct" AI than ChatGPT. Meanwhile, the Chinese government has promulgated rules to ensure generative AI reflects fundamental socialist values.

Marchand cautions against "fantasizing too much about the massive manipulation that this medium could represent," noting that a conversational robot that always goes in the same direction may interest fewer people.