Super-intelligent AI will soon be difficult to tell apart from humans – expert warns of ‘pain’ but promises big perk
ARTIFICIAL intelligence chatbots will soon be so good that you’ll struggle to tell them apart from humans.
A top security expert told The Sun that despite the risks of convincing AI, this is actually a good thing – although he warned of a period of “pain”.
Chatbots like OpenAI‘s ChatGPT or Google‘s Gemini can make boring tasks much easier.
You can use them to get work done more quickly – and companies are rolling them out to deal with simple (and sometimes even complicated) customer support problems.
Part of this involves artificial intelligence chatbots speaking in a more natural humanlike way, and telling them apart from real people will be a big challenge.
“There is absolute certainty that chatbots and AI-enabled chatbots are going to get better and better over time,” said cybersecurity expert Roger A. Grimes, speaking to The Sun.
“Yes, it will become more difficult over time to tell if you are talking to a human or a chatbot.
“And as it becomes more difficult to tell, that’s a good thing.”
Many experts have warned of the security dangers of advanced AI – and how cyber-crooks could abuse them to scam innocent victims.
But Roger, a defence evangelist at security firm KnowBe4, told The Sun that he thinks chatbots becoming more humanlike is generally a good thing.
“It means we are getting the help we desired and needed…faster,” he said.
“Who doesn’t want that? The whole reason most of us want a human to help us today versus a chatbot is that the chatbot isn’t all that helpful.
“It just wastes our time and causes additional frustration.
“But as they become more popular, it means they will be more capable of solving more of our problems in a timely manner.
“If a chatbot is going to solve my problem and I don’t have to wait longer for a human, I’d prefer the chatbot.”
TRICKY TECH
The problem for people using chatbots is that they’re still not quite there.
That will leave humans free to handle truly unusual scenarios.
Roger A. Grimes
Chatbots are known to “hallucinate” and serve up totally false information.
And they can sometimes totally misunderstand what you’re asking.
These situations are much more rare with humans – and although chatbots are getting better, Roger said it’s causing “pains” right now.
“The problem is that we are feeling the pains of the transition period where the chatbots aren’t so awesome,” Roger told The Sun.
SPOT THE SIGNS OF A CHATBOT
Here’s what The Sun’s Head of Technology and Science Sean Keach has to say…
One day, it very well might be impossible to tell a chatbot apart from a human.
But for now, there are plenty of giveaways that you’re speaking to an AI.
Often they’ll write very formally in a way that a human typically wouldn’t.
They’ll also reply extremely quickly. Maybe you’ve been ghosted by a mate on text. AI won’t do that. Unless you’re really boring.
But really, watch out for ultra-fast response times, especially if they’re churning out massive paragraphs at breakneck speed.
You might find they use repetitive phrasing, or strange terms you wouldn’t normally expect to see (the word “delve” is classic).
They might make strange mistakes (like the classic chatbot error of miscounting the number of “r”s in the word strawberry) that a human probably wouldn’t.
And the chats might not flow like they would with a real human either.
If you want to test a chatbot, trying throwing it off with off-topic questions or asking for personal anecdotes.
With these tricks, you should hopefully avoid being fooled by brainy AI bots.
“Especially when dealing with less popular requests and problems.
“But over time, the chatbots will get trained to handle less common problems and scenarios and get us to the right solution quicker.
“That will leave humans free to handle truly unusual scenarios that require a more intelligent hand.”