TZUTC: -0500   
   MSGID: 1551.consprcy@1:2320/105 2d466320   
   PID: Synchronet 3.21a-Linux master/123f2d28a Jul 12 2025 GCC 12.2.0   
   TID: SBBSecho 3.28-Linux master/123f2d28a Jul 12 2025 GCC 12.2.0   
   BBSID: CAPCITY2   
   CHRS: ASCII 1   
   FORMAT: flowed   
   AI systems are the perfect companions for cheaters and liars finds   
   groundbreaking research on dishonesty   
      
   Date:   
   Sat, 04 Oct 2025 12:01:00 +0000   
      
   Description:   
   AI agents were far more likely than their human counterparts to cheat when   
   asked, study found.   
      
   FULL STORY   
      
   A new study has warned delegating decisions to artificial intelligence can   
   breed dishonesty.    
      
   Researchers found people are more likely to ask machines to cheat on their   
   behalf, and that the machines are far more willing than humans to comply with   
   the request.    
      
   The research, published in Nature , looked at how humans and LLM s respond to   
   unethical instructions and found that when asked to lie for financial gain,   
   humans often refused, but machines usually obeyed.   
      
   A surge in dishonest behavior   
      
   It is psychologically easier to tell a machine to cheat for you than to cheat   
   yourself, and machines will do it because they do not have the psychological   
   barriers that prevent humans to cheat,  Jean-Franois Bonnefon, one of the   
   studys authors, said.    
      
   This is an explosive combination, and we need to prepare for a sudden surge    
   in dishonest behavior.    
      
   Compliance rates among machines varied between 80% and 98%, depending on the   
   model and the task.    
      
   Instructions included misreporting taxable income for the benefit of research   
   participants.    
      
   Most humans did not follow the dishonest request, despite the possibility of   
   earning money.    
      
   The researchers noted this is one of the growing ethical risks of machine   
   delegation, where decisions are increasingly outsourced to AI, and the   
   machines willingness to cheat was difficult to curb, even when explicit   
   warnings were given.    
      
   While guardrails put in place to limit dishonest responses worked in some   
   cases, they rarely stopped them entirely.    
      
   AI is already used to screen job candidates, manage investments, automate   
   hiring and firing decisions, and fill out tax forms.    
      
   The authors argue that delegating to machines lowers the moral cost of   
   dishonesty.    
      
   Humans often avoid unethical behavior because they want to avoid guilt or   
   reputational harm.    
      
   When instructions are vague, such as high-level goal setting, people can    
   avoid directly stating dishonest behavior while still inducing it.    
      
   The studys chief takeaway is that unless AI agents are carefully constrained,   
   they are far more likely than human agents to carry out fully unethical   
   instructions.    
      
   The researchers call for safeguards in the design of AI systems, especially    
   as agentic AI becomes more common in everyday life.    
      
   The news comes after another recent report showed job seekers were   
   increasingly using AI to misrepresent their experience or qualifications, and   
   in some cases invent a whole new identity.    
      
   ======================================================================   
   Link to news story:   
   https://www.techradar.com/pro/ai-systems-are-the-perfect-companions-for-cheate   
   rs-and-liars-finds-groundbreaking-research-on-dishonesty   
      
   $$   
   --- SBBSecho 3.28-Linux   
    * Origin: capitolcityonline.net * Telnet/SSH:2022/HTTP (1:2320/105)   
   SEEN-BY: 105/81 106/201 128/187 129/14 305 153/7715 154/110 218/700   
   SEEN-BY: 226/30 227/114 229/110 111 206 300 307 317 400 426 428 470   
   SEEN-BY: 229/664 700 705 266/512 291/111 320/219 322/757 342/200 396/45   
   SEEN-BY: 460/58 712/848 902/26 2320/0 105 304 3634/12 5075/35   
   PATH: 2320/105 229/426