Public to be warned about AI scams?

Not sure I follow this.


It follows the development of advanced AI technologies such as GPT4, which can convincingly mimic a human in conversation – including one example where it falsely told a human user that it was not a robot in order to convince them to complete a task on its behalf.


He said AI could be used by bad actors to “mechanise and automate” social engineering scams such as romance fraud. Such frauds are usually perpetrated by an individual human, but AI technologies could enable them to become more widespread. “You can be having a conversation with someone by text or by social media who convincingly appears to be a real person,” he said.

If I get a call from someone asking for my pin number - it doesn't really matter to me whether it's a real person or a ChatGPT pretending to be a real person?

The fact that it's not a real person, doesn't make it any less scammy?

What am I missing here?

Brendan

Brendan Burgess said:

What am I missing here?

Click to expand...

The need for newspapers to publish clickbait articles and sow, arguably overblown, fear in the public over the latest supposed threat to humanity?

Its still a scam, but now can be automated.

faster responses wider reach, cheap (from scammers point of view) and can glean info on target quickly.

Brendan Burgess said:

What am I missing here?

Click to expand...

They no longer have to engage a call center to make calls. They can reach many multiples of potential victims than they could previously if it's fully automated. And, the AI can learn the most effective way to succeed in parting a victim from their cash.

AI can speak almost any language. With a few pieces of personal information (name, age, location) it can find your social media accounts, company registration records, Leinster Leader article about your aunt winning the ploughing championship, etc. It can then compose a bespoke story to trap you. It can calculate how likely you are to fall for the con depending on how long it takes you to reply, what you say, etc. and it will adjust its response accordingly. They only need a tiny success rate to be profitable.

I figured I’d cut to the chase and just ask the AI itself whether this is something to be concerned about or not -

Artificial intelligence can be used to make scams and identity theft quicker and easier in several ways:

1. Automated phishing emails: Scammers can use AI to generate phishing emails that look like they were sent from a legitimate source. These emails can be personalized to the victim's interests, location, and other information obtained from their online profiles, making them more convincing and increasing the likelihood of the victim clicking on a malicious link or entering their login credentials.
2. Voice phishing: AI-powered voice cloning technology can be used to mimic the voices of trusted individuals, such as family members, coworkers, or even public figures, to trick victims into revealing personal information or transferring funds. These voice phishing attacks can be highly convincing, making it difficult for the victim to detect that the caller is not who they claim to be.
3. Data mining: AI algorithms can be used to scrape large amounts of public data from social media platforms, public databases, and other sources to build detailed profiles of individuals, including their personal information, interests, and behaviors. This information can be used to craft personalized phishing attacks or to launch identity theft schemes.
4. Fraud detection: AI can be used to bypass fraud detection systems by generating patterns that mimic legitimate transactions or by finding vulnerabilities in the system that can be exploited to steal funds or personal information.

Overall, while AI can be used for many positive purposes, it is important to be aware of its potential for misuse and to take steps to protect your personal information and financial assets from scams and identity theft.

Click to expand...

FWIW to another question I asked it, it pointed out that Microsoft recently demonstrated technology to mimick a voice after hearing just 3 seconds of speaking. Pair that with a convincing language model like ChatGPT and you’re no longer getting a call in broken English but what sounds like a call from your spouse or sibling asking you to urgently approve a bank transaction or remind them of a PIN. The scammers cost has just gone through the floor as they don’t need to hire people and their success rate has gone through the ceiling, not a happy mix for us.

If you thought the unintended consequences from social media were bad, wait until you get a load of AI.