Although generative artificial intelligence can be useful for journalists in the fact-checking process, the technology’s limitations pose challenges, especially in countries with less widely spoken languages.
Artificial intelligence helps fact-checkers, but the technology is imperfect
Although generative artificial intelligence can be useful for journalists in the fact-checking process, the technology's limitations pose challenges, especially in countries with less widely spoken languages.
Can generative AI help journalists debunk false facts? Reuters journalist Gretel Kahn looked for answers to this question, reviewing the practices of fact-checking with the help of AI in Norway, Georgia and Ghana. AI-powered tools are speeding up the work of fact-checkers, but they are not yet a panacea, according to Kahn’s interviewees. The limitations of technology become apparent when it is applied in non-Western countries or countries whose national languages are not widely spoken in the world. For example, Georgia-based fact-checking platform MythDetector is experimenting with generative AI tools to help detect falsehoods spreading on online platforms. However, the tools they use are not trained in the Georgian language, so there are a lot of mistakes. For this reason, fact-checking teams must review selected content.
