HUMANS will need to change their behaviors to avoid falling victim to deceptive chatbots.
A cyber-expert has told The U.S. Sun how you must avoid putting too much trust in artificial intelligence.
AI chatbots are increasingly popular, with tens of millions of people flocking to apps like OpenAI's ChatGPT and Google Gemini.
These chatbots use large language models that allow them to speak to you just like a human.
In fact, a study recently claimed that OpenAI's GPT-4 model had passed the Turing test – meaning humans couldn't reliably tell it apart from a real person.
What Ola and James Jordan really ate and did to shed 7st
We spoke to cyber-expert Adam Pilton, who warned that the humanlike way chatbots talk makes them much more capable of deceiving us.
"It feels as though it would be easier to be drawn in by the conversational nature of a chatbot, versus perhaps a deceptive website or search result," Pilton said.
He should know. He is a Cyber Security Consultant at CyberSmart and a former Detective Sergeant investigating cybercrime.
"As humans, we build trust where we potentially see a relationship and it's a lot easier and understandable to be able to build a relationship with a chatbot compared to a website," he continued.
"A website doesn't respond to our specific requests whereas with the chatbot we feel like we're building a relationship because we can ask it specific questions.
"And the answer it gives us is tailored to specifically address that question.
"In this modern digital world we are living in a key skill that will now be the verification of information, we cannot simply trust what we are first told."
SNEAKY SPEAKERS
Earlier this year, scientists revealed how AI had mastered the art of "deception" – and learned it on their own.
And chatbots are even capable of cheating and manipulating humans.
Spotting the signs that a chatbot is trying to trick you is important.
I'm a 'time traveler' - the 'worst case scenario that could kill us all'
But Adam warned that we must now adopt a "new way of life" where we don't trust AI chatbots – and instead verify what they tell us elsewhere.
"Disinformation and simply incorrect information is going to be an increasing problem for society and democracies around the world as we continue to evolve in this digital world," Adam told The U.S. Sun.
"As such the verification of information is going to be a common requirement and the use of chatbots is no different.
"We can no longer depend upon a single source of information verification from multiple trusted sources is now a way of life."
SHARE CARE
Chatbots will only become more popular over time as their capabilities grow.
But there are many risks, including giving too much of your own information over to them.
Experts recently warned The U.S. Sun about the importance of not telling an AI too much about yourself.
They've even been described as a "treasure trove" for criminals looking to find out info about victims.
Used safely, chatbots can be hugely helpful – but be careful not to tell them too much, and don't trust everything they say.
