Introduction

For decades, search engines like Google and Bing played a fundamental role in "providing knowledge," offering users vast amounts of information while leaving them the freedom to choose and analyze. However, with the emergence of artificial intelligence (AI) and applications like ChatGPT, Bard, and Copilot, the paradigm has shifted from "providing knowledge" to "shaping knowledge." Users are no longer independent researchers but rather participants in a process that may deliver pre-packaged information—often inaccurate or biased.

Providing Knowledge: The Era of Search Engines

In the past, search engines acted as neutral intermediaries, indexing and displaying information based on algorithms that determined relevance to the user's query. The key advantages of these platforms were:

1.     Relative Transparency: Users could see multiple sources and choose what suited them.

2.     Freedom of Choice: Users decided which sources to trust and which to ignore.

3.     Diversity: Different perspectives on the same topic were presented.

This stage was based on the principle that users could distinguish between truth and falsehood and had the necessary tools to analyze information independently.

Shaping Knowledge: The Era of AI

With the rise of generative AI (like ChatGPT), the equation has changed. These tools no longer just provide information—they shape and frame it in ways that may hide biases or errors. Key changes include:

1.     Pre-Packaged Answers: Instead of listing multiple sources, AI delivers a single, ready-made response, limiting the diversity of viewpoints.

2.     Hidden Bias: Language models are trained on potentially biased data, meaning their answers may reflect the biases of developers or training sources.

3.     The End of Transparency: AI does not always disclose its sources, making it harder for users to verify information.

4.     Weakening Critical Thinking: Reliance on ready-made answers may reduce users' ability to research and analyze independently.

The User: From Decision-Maker to Passive Recipient?

Previously, users made decisions based on diverse information. Today, some AI applications present answers as if they were the "absolute truth," making users more susceptible to intellectual dependence. Even worse, some of these models:

·       Oversimplify information, omitting crucial details.

·       Blur facts and opinions without clear distinction.

·       Deliver incorrect answers with high confidence (a phenomenon known as hallucination).

Challenges and Risks

This shift from "providing knowledge" to "shaping it" poses serious challenges, including:

·       Manipulation of Public Opinion: AI can be used to push certain narratives without allowing counterarguments.

·       Decline in Independent Research: Users may rely solely on AI, reducing the diversity of knowledge.

·       Accountability Issues: Who is responsible if AI provides false information leading to poor decisions?

How Can We Address This?

To mitigate these risks, we must:

1.     Enhance Transparency: AI platforms should disclose their information sources and limitations.

2.     Promote Digital Literacy: Educate users on verifying information and avoiding over-reliance on AI.

3.     Regulate AI Use: Establish ethical and legal safeguards to prevent misinformation.

Conclusion

AI has transformed how we access knowledge, shifting from a model of "provision" to one of "shaping." This makes users more vulnerable to algorithmic biases. If we do not act cautiously, we risk transitioning from a knowledge-seeking society to one that is fed pre-packaged information. Therefore, it is essential to develop mechanisms that preserve intellectual independence and ensure AI remains a tool for knowledge—not its master.