To keep up with the rapidly advancing AI landscape, consumer-facing solutions like Midjourney and ChatGPT have been created to produce amazing images and texts quickly in response to natural language commands. They are being used far and wide, from web searches to children’s literature.
AI applications are being misused for evil purposes, such as spreading malware. This includes scam emails — these typically have glaring mistakes in terms of grammar and spelling. However, AI models have become so advanced that they no longer make such errors – a fact pointed out in a recent Europol advisory report.
Social engineering, used to access passwords, financial information, or other sensitive data, has become easier with automated systems that generate text persuasive enough to dupe users. This text can be modified and adapted for different audiences with minimal effort and maximum results regarding phishing attacks or other security threats.
OpenAI, who created ChatGPT, implemented certain safeguards: if asked to “write malware” or a “phishing email,” it will follow strict ethical guidelines and refuse to do so. It is programmed not to engage in any malicious activities related to writing or assisting with creating malware.
Despite the safeguards installed to prevent ChatGPT from crafting malicious codes, it can still be used to craft emails by cybercriminals. There are signs that people are trying to circumvent these security features.
Using ChatGPT as an example, it’s not hard to see how criminal groups can utilize large language models and similar tools to craft more believable scams. Text is not the only targeted area; audio and video forgery is becoming increasingly common.
AI bots are extremely effective when attempting to scam people, using trust and the appearance of authenticity. When a boss demands an urgent report, tech support insists on installing a security patch, or a bank alerts customers to an issue that needs replying to – all of these could be fake.
Natural, tailored, and rapid content production on demand – this is the capability that modern AI-generated media has to offer. Audiences can enjoy audio, video, and text materials; generated quickly and perfectly adapted to them.
We can fight back against AI-powered threats; there is still hope! To reduce the chances of being scammed, the same precautions that have long been used should be taken – nothing new is required.
Protect Yourself From AI-powered Scams – How To Guard Against Them
Two specific security risks arise from using AI-based technologies like ChatGPT or Midjourney. These could include tricks wherein someone could persuade you to install unnecessary browser plugins, pay for services you don’t require, or use bogus applications which seem legitimate.
To prevent such mistakes, it is crucial to stay informed of the developments in AI services such as ChatGPT. Bear in mind there is no authorized app available, and the tool can only be accessed through its web version. Therefore, always find your information in the source first!
When working with any of these apps and derivatives, you must follow the same rules: Check their past, review companies related to them, and the associated reviews. By doing so, users ensure the safety of installing any new app.
The other danger is AI that produces text, audio, or video content that could be realistic enough to deceive the listener. For instance, one employee fell for a voice clip supposedly from their chief executive asking for an urgent money transfer. This shows the potential harm caused by AI-generated material appearing convincing.
Use caution when performing a task out of the ordinary; as technology has evolved, criminals may still use time limits and urgent requests to manipulate you. Be aware of any red flags; double-check emails or phone calls with different methods, and take your time on suspicious or unfamiliar requests.
Never follow suspicious links in texts and emails, particularly if you are asked to log in somewhere. To be safe, if you receive a message from your bank, for example, go directly to the bank’s website in your browser to sign in instead of any link they provided.
Maintaining your OSes, apps, and browsers up-to-date is critical; luckily, it often occurs automatically. This protects you against many phishing and scam attacks — whether created by AI or not — since more contemporary browsers have improved security features that recognize these attempts to deceive.
There is no definitive method to detect AI texts, audio, or videos. However, specific clues like indistinct image collages and generic sentences could be identified. Fraudsters may have obtained details concerning their life or job from a source, but they probably do not completely comprehend your personal affairs.
The recent emergence of AI services has made it all the more important to exercise caution and remain skeptical – a truth that was also applicable before their advent. Just like the imagined face-morphing masks from Mission: Impossible, one should be sure of whom they’re dealing with before disclosing any information.
Bot developers are responsible for prioritizing user safety and taking steps to prevent the spread of malware through their bots. This includes rigorous testing and monitoring of their code and educating users about potential risks and best practices for safe bot usage.
By taking these steps and staying informed about the latest threats and security measures, users can continue to enjoy the benefits of chatbots without putting themselves at risk of malware and other security threats.
Source: WIRED
