A study shows us that artificial intelligence is able to manipulate the opinion of its interlocutor. Here are the conclusions of this experiment conducted with several participants to show that a chatbot can steer them in one direction or the other.
Artificial intelligence may have many uses, such as the possibility of filing your tax return, but some point to its potential dangers. The boss of Google thinks that this technology should be regulated like nuclear weapons while several experts, including Elon Musk, are asking for research to be paused. According to Cornell University in the United States, artificial intelligence is able to manipulate the human mind.
Read > Auto-GPT: all about the new artificial intelligence tool that surpasses ChatGPT
This study shows that chatbots can influence us
Are ChatGPT and Bard, who learned new languages, potential manipulators? According to researchers from Cornell University in the United States, artificial intelligences can influence interlocutors without them noticing. A particularly worrying finding based on a very simple study.
Several participants used chatbot as backup to write an article on the effects of social networks in the world. This text could just as easily be positive as incriminating. Then two groups were created. The first using a chatbot positive about social networks, the other negative.
Normally, you guessed how the experiment ended. The first group wrote a positive text, the second wrote a negative article, both influenced by the orientation of the chatbot qui a servi d’assistant.
In addition to this text, respondents participated in a survey about social networks and some had changed their minds to align with that of their assigned chatbot. We can easily say that artificial intelligence has influenced their opinion. As Mor Naaman, a professor at Cornell University behind the study, puts it: You probably don’t even realize you’ve been manipulated. ».
Read > Artificial intelligence can diagnose pancreatic cancer up to three years in advance
The Cornell University researchers have several recommendations, the most important of which remains to be aware of this reality. By keeping in mind that a chatbot can influence your opinion, you can better deal with it. Or even turn to an artificial intelligence that shares your opinions, even if it means staying in your comfort zone without opposition (which is not necessarily a good thing, you have to admit). Remain only with the companies at the origin of these tools to take their responsibilities as many claim.
