The development of conversational agents in AI language models differs from the contributory model that has shaped Internet content. Users cannot directly publish content, critique query results, or evaluate products recommended by chatbots. The training procedure for language models is controlled by private companies with commercial interests, hindering the creation of a truly participatory model. AI is being used to limit users’ freedom of expression, with censorship policies in place.
A conversational agent should provide a balanced view of arguments in debates without taking a definitive stance or suppressing viewpoints. This impartiality should extend to commercial product guides to ensure competition law compliance and quality advice. The inclusion of user contributions is crucial for the qualitative performance of AI models.
In order to enhance the qualitative efficiency of new language models, institutional actors, associations, and civil society must promote AI software that respects fundamental freedoms and network neutrality. Political will is needed to integrate public participation in training algorithms. Real-time algorithmic self-training could be used in courts and parliamentary assemblies.
The future of AI technology lies in dynamic AI 2.0 models that incorporate user contributions for qualitative evaluation. However, obstacles such as censorship, cultural stereotypes in training data, and economic biases of software developers hinder the qualitative efficiency of language models. The political and psychological choices of AI designers play a significant role in determining the quality of generative intelligence software and their relationship to collective intelligence.
Source link
Source link: https://hackernoon.com/towards-network-neutral-ai-and-open-dialogue