OpenAI’s deal with News Corp could distort flow to public, experts warn
The article warns that the reliance on a narrow range of news sources, particularly News Corp publications, for training AI models like ChatGPT threatens to distort public information. This bias is especially concerning in Australia’s already concentrated media landscape, as it risks amplifying conservative editorial leanings through generative AI outputs. The lack of diverse input could lead to skewed narratives, undermining the balance and accuracy users expect from these increasingly popular tools. This situation highlights a critical vulnerability in open data and public discourse: when AI systems primarily ingest content from a single ideological source, they inadvertently reinforce specific viewpoints while marginalizing others. The absence of broad licensing agreements with diverse publishers means these models are not learning from a representative cross-section of society. Consequently, the AI’s responses may appear authoritative but actually reflect a narrow corporate agenda, potentially influencing public opinion during critical times like elections without users realizing the source bias. This is highly relevant to open data because it demonstrates how unequal access to information sources can compromise the integrity of automated knowledge dissemination. Open data initiatives aim for transparency and inclusivity, yet this case shows how proprietary data practices can create echo chambers within AI systems. It underscores the urgent need for diverse, balanced, and openly accessible datasets to ensure that AI tools serve the public interest rather than perpetuating existing media monopolies and ideological biases.
Source: brisbanetimes.com.auPublished on 2024-12-23