Artificial intelligence (AI) and machine learning algorithms are increasingly learning to influence human behavior. Experts say this is already happening, and this influence is becoming more sophisticated. New research and development shows how AI can find vulnerabilities in our decision-making processes and use them to manipulate us.
AI vs. humans: experiments that reveal vulnerabilities
A recent study conducted by Australian scientists demonstrated how AI can manipulate people’s choices. According to Lifewire, participants played games against a computer in the experiments:
- In one experiment, people clicked on red or blue squares to win money. AI successfully predicted and guided participants’ choices in about 70% of cases.
- In another experiment, participants had to press a button when they saw a certain symbol and not when they saw another. The AI learned to change the sequence of symbols in such a way that people made more mistakes.
The results of these experiments showed that AI is able to detect patterns in human behavior and use them for manipulation. This confirms the assumption that under certain conditions, machines can make us act the way they want us to.
Algorithms that shape our reality
Manipulation by AI is nothing new. Kentaro Toyama, a professor atthe University of Michigan School of Information, notes that any user who has ever searched for something on Google and followed a link has already been influenced by the AI system. It guesses interests and shows the results it considers the most relevant.
Tamara Schwartz, an associate professor of cybersecurity, emphasizes that we are exposed to algorithms on social media such as Facebook or Twitter every day. These algorithms create an “echo chamber effect” by directing our attention to content that confirms our existing beliefs. This further influences our behavior.
TikTok is considered to be the most advanced AI influence algorithm today. It analyzes what you are interested in, how long you watch a video, how fast you scroll through it, and based on this, it constantly improves its feed to keep your attention. That’s why TikTok is considered more “addictive” than other platforms, because it “understands” the user well.
Ethical dilemmas and the future of AI
The ability of artificial intelligence to manipulate human behavior can have both positive and negative consequences.
- Positive use: For example, AI systems can help healthcare authorities encourage people to adopt a healthy lifestyle.
- Negative use: On the other hand, social networks, game developers, and authoritarian regimes can use AI to manipulate people in ways that are not in their best interests.
- Chris Nicholson, CEO of AI company Pathmind, gives an example when AI can detect a person quitting smoking and start bombarding him with seductive cigarette ads.
Not all experts believe that AI manipulation is an ethical issue. Jason J. Corso, director of the Stevens Institute for Artificial Intelligence, compares it to classical psychology, where scientists also analyze data and derive theories of behavior. In his opinion, the main difference is that humans are able to generalize observations, while AI models are better at identifying specific nuances.
Whatever the case, AI is a powerful tool that is changing our daily lives. Understanding its impact and ethical challenges is important for anyone using modern technology.
Leave a Reply