Threats Modern AI Brought Us

Written by an AI user for AI users

Bill Gates said:

The development of AI is as fundamental as the creation of the microprocessor, the personal computer, the Internet, and the mobile phone. It will change the way people work, learn, travel, get health care, and communicate with each other. 

And we can’t disagree with it. 

However, AI is a revolutionary tool it brought us many threats. I won’t talk about machine wars, as you can read about them in fiction books and Twitter threads. We’ll check more down-to-earth threats people are likely to experience tomorrow. 

So let’s cut it short and go straight to the list:

• A brilliant tool for fakes
• AI can lie
• Copyright question
• Accounts hijacking
• Malware distribution

You may also like AI Trends of 2023. 

A brilliant tool for fakes

With AI, everything becomes unreliable, as you can fake easily. 

• Voice replication

AI can replicate any voice and sound like your friend or Paul McCartney. 

Just check this post: 

source: Instagram​

• Generating photorealistic images

As you know, Midjourney is able to create images from text descriptions. With the AI revolution, it also has improved significantly. The latest version can produce photorealistic images.

source: Twitter​

source: Twitter​

It seems that everyone saw this puffer coat look, but only a few knew it was fake and made by AI.

source: Twitter​

The wisdom says, "To fight AI, create another AI". We need one which will identify photos, texts, and other content created by artificial intelligence. 

But there is a problem: most such models are learned from a particular AI generator and can't identify fakes created by other algorithms. There always will be a race between these types of AIs as they will be taught and improved by one another. 

But now you can already check the photos here.

AI can lie

Ai can use false information when answering the users. For example, one lawyer asked an AI chatbot to create a list of legal scholars who committed sexual harassment. ChatGPT gave him a list with Mr. Turley’s name. It wasn’t true, and Chat falsely accused him of making sexual suggestions and trying to touch one of his students when they were on a trip to Alaska. It referred to a post from 2018. The George Washington University professor claimed no such article existed.

There’s one more famous case of AI lying. ChatGPT said that an Australian mayor had been imprisoned for bribery. Brian Hood, the mayor of Hepburn Shire, claimed that it was falsely generated by an AI chatbot. His lawyers sent a letter to OpenAI, giving the company 28 days to fix the errors about Mr. Hood.

AI can provide you with 100 ways of fact-checking. However, it doesn’t use any of them. This can spread false information, as humans also don’t check everything AI says

Copyright question

I want to clarify this question to everyone using AIs in their work, especially creative ones, as there is a chance that the works you create now with AI won’t be protected by copyright. 

The ChatGPT was mentioned as the author or co-author of 200+ books in Amazon’s bookstore. But we understand that the real number of books created by AI could be much higher as Amazon doesn’t require users to mention it. 

More likely, all your works and texts that are created just with AI prompts won’t actually be yours. Why do I think so?

The comic book Zarya by Kristina Kashtanova featured pictures created by feeding text prompts to Midjourney.

The US Copyright Office sent the letter (PDF) to her lawyer by Robert Kasunic, the associate Register of Copyrights. They decided that Kashtanova “is the author of the Work’s text as well as the selection, coordination, and arrangement of the Work’s written and visual elements”. The images themselves, however, “are not the product of human authorship,” and the registration originally granted for them has been canceled. 

Actually, in most copyright acts, the subject of the copyright can be only a human. So the authors of those 200+ books placed on Amazon shouldn’t now name ChatGPT their co-author. It would be more logical in the context of the available acts to name the founders of ChatGPT as co-authors, as they’re humans, at least. However, we’re waiting for the new version of acts to regulate this question. USA Copyright Office is preparing them. 

In Australia, works generated by AI can’t be protected by copyright now and so can be used by anyone. Lawyers now say that if the work was generated by AI and then significantly changed or improved by humans, he/she can become the author and get the copyright. 

I want to highlight that we’re ONLY WAITING FOR A REGULATORY SYSTEM, and the things above are only facts and thoughts.

Accounts hijacking

Google has removed a ChatGPT plugin from the Chrome web store after it was discovered to be collecting Facebook cookies. Before this move, 9000 individual accounts were affected.

The malicious ‘Chat GPT’ extension was based on the original open-source project and had a similar name to the legitimate ‘ChatGPT for Google’ extension. The malicious actors behind the fraud changed a few lines to the original code.

The false extension appears and behaves in the same way as the actual ChatGPT extension, making it difficult for users to identify. Furthermore, its position on the Chrome online shop meant, unfortunately, it had a significant number of downloads.

Following installation, the extension can block Facebook cookies, which attackers can then exploit to modify login data, potentially utilizing the stolen accounts to promote fraud and other propaganda while shutting out the original users.

Malware distribution

Cloudsek cybersecurity researchers revealed that the amount of AI-generated YouTube videos infected with and spreading malware has increased by 200-300% month on month.

These movies, however, contain links to malware that steals information, commonly known as info stealers, such as Vidar, RedLine, and Raccoon. Links are disguised by employing URL shorteners (the most prevalent being telegra.ph, bit.ly, and cutt.ly), being linked to genuine file hosting services such as mediafire.com, and using URLs that directly download the malicious zip file.

So...

We know AI as a revolutionary thing, but together with a new tech era, it brings numerous threats we must consider. 

• Creating fakes is now easier than ever. AI can copy the voice, make a realistic photo and write anything from email to a book. 
• Not everything ChatGPT answers is true. Double-check its texts. 
• Remember to use AI as a helper, don’t replace yourself with it, especially when doing work that could be protected with copyright in the future. As you more likely won’t have an opportunity to protect it with the copyright of it generated by artificial intelligence. 
• Remember all online safety rules not to let AIs affect you or steal some info. Better double your protection.

    We should mind threats, but we should also use the AI opportunities. CGS-team knows how to boost your business with AI API integration service.