Why do we trust chatbot answers without checking the sources?

Although chatbots provide smooth and confident answers, they often hide disagreements between sources and reduce our motivation to verify information, which can lead to the spread of misinformation that users feel fully aware of.

Read more

A first-of-its-kind lawsuit against Grok over child pornography

Studies indicate that about 900 million people use ChatGPT weekly, and that 10% of people around the world use chatbots to get news, with the percentage rising to 16% among those under 35 years of age.

But only 42% of those users check sources by clicking on links, while Pew Research Center research reports that having an AI summary above search results reduces clicking on actual sources by nearly half.

Why do we trust it without verification?

This is due to three main factors:

First: Ease of reading makes the information seem more correct, as the immediate and clear responses from the robot put pressure on our psychological tendencies to accept it as truth.

Read more

Is artificial intelligence making us more like robots?

Second: The comfort of getting a single answer, even if it is inaccurate, overcomes the discomfort resulting from inconsistency between different sources, which increases our confidence in a single answer even when multiple answers are more accurate.

Third: What is called “Machine Heuristic”, which is our tendency to assume that machines are objective and do not have special agendas, which helps to overcome the defenses that may be raised by human sources.

The biggest problem: unreliable answers

Despite this confidence, a study conducted by the BBC in cooperation with 18 countries and 14 languages showed that almost half of chatbots’ answers to news contain at least one major problem, and the most common were errors in sources, with misleading, missing, outdated, or even completely fabricated information.

Read more

Study: Increasing cases of artificial intelligence being out of users’ control

With fewer people checking sources, this results in a misinformed public that may think itself fully informed without realizing that its information is incorrect.

Solution: Add a deliberate habit to the automatic habit

Experts recommend taking two steps: First, awareness of the psychological tendencies that make us accept claims uncritically. Secondly, treating the robot’s answer as the beginning of understanding the topic, not the end, by requesting and verifying sources, and examining the topic from different angles and conflicting sources. The truth is often in what we gather from multiple angles, not in the one line that the chatbot gives us.