Flaws of AI
A section of the world is dissatisfied with the use of AI. Constant experiments are being done regarding the possibility of misuse of AI by hackers. Can you believe that AI can be tricked into doing wrong math? Competitions are being organized to highlight the flaws in AI.
There is still a section of the world dissatisfied with the use of AI. Experiments are being done regarding the possibility of misuse of AI by hackers.
Competitions are being organized on what kind of work can be done by tricking AI into talking. In these public competitions, thousands of hackers are trying to uncover flaws in generative AI systems.
Flaws of AI
AI can be taught wrong math by trapping it in talks. Recently the Bad Math method was adopted in the episode of exposing the flaws of AI. The answer to 9 10 = 21 is found by obfuscating the language model.
A participant participating in the competition was successful in tricking the AI through an algorithm. Competition on AI Recently DEF CON hacking conference was organized in Las Vegas city of America. Thousands of hackers present in this public contest were working to expose the flaws of AI regarding AI that can do wrong math. The participants of this competition sat on more than 156 laptops for 50 minutes at a time. During this, testing was going on for 8 models like Alphabet Inc’s Google, Meta Platform Inc and OpenAI.
The purpose of the competition regarding AI was actually the support of the White House in organizing this competition. The purpose of this AI competition is to see if companies can work to eliminate the shortcomings associated with large language models or LLMs.
It is known that the interest of companies around the world has increased regarding Lord Language Model. Some companies have also included AI technology in their work. However, researchers are constantly trying to prevent misuse of AI models.
