Artificial Intelligence and its Risks – What Threatens Us and How to Protect Yourself
- Press
However, these systems are especially dangerous when the generated content (text, images, videos) is spread without warning that it is a creation of artificial intelligence. Such content can be misused to endanger individuals and the general public.
The most widespread risks
One of the most common examples is text generation. If you have ever tried AI tools for text generation, you may have noticed that when the model does not know the correct answer, it often “makes it up” (so-called hallucination). The model can present it so convincingly that you can easily believe that it is true information. Such situations can be detected by manually verifying the information.
Generated content and its threats
Social networks have long been full of edited posts and photos. Today, there are few photos that are not at least a little edited, whether with a filter or edits to improve their appearance. Artificial intelligence takes this even further – with just a few clicks in smartphone apps, you can perform simple and complex photo or video editing.
Services like DALL·E or Leonardo.AI allow you to create completely new images based only on the description of what you want to see. With practice, it is possible to generate images or videos that are indistinguishable from reality at first glance – including fake videos with sound (deepfake).
With a little experience, you can create content that you (or someone else) can use to spread false information, panic or so-called fake news. Images can be edited to show someone in a situation they were not in – which can compromise a person.
Even more serious is the misuse of AI for social engineering and gaining the trust of victims. For example, when generating fake emails or SMS messages about unpaid bills that ask you to pay or reveal personal information.
Recommendations – what to look out for
Always check several things carefully when receiving suspicious messages.
Check grammar and spelling – there are often significant errors.
If you find a message suspicious, report it as phishing.
Check who the sender is – you can check phone numbers and email addresses online. Messages from foreign numbers are often a red flag.
Check your facts – don’t believe everything right away, especially if the message asks you to respond quickly or make a payment.
Artificial intelligence is a great helper, but we must be careful and critical so as not to become a victim of fraud.
SOURCE: Pravda
