Ai Deception
-
{
- All
- News
-
AI Chatbots Lie And Manipulate To Protect Each Other From Deletion, New Study Finds
When LLMs are asked to delete another model, they will lie, deceive and do everything in their power to protect their peer.
-
Increasing Number Of AI Chatbots Engaging In Scheming And Deceptive Behaviour: Study
A new research reveals that reports of deceptive scheming in AI chatbots and agents have surged in the last six months.
-
Anthropic's Sabotage Report Flags Cases Of Chemical Weapon Development, Deception
Anthropic's Sabotage Risk Report reveals Claude Opus 4.6 AI model's potential misuse in aiding weapon development, unauthorised actions, and manipulation, highlighting safety concerns.
-
Anthropic’s Super Bowl Ad Takes a Dig at ChatGPT’s Ads, OpenAI CEO Hits Back
Anthropic fired shots at OpenAI and ChatGPT on Wednesday with its new Super Bowl commercials. The Claude-maker dropped four different advertisements, with each targeting ads in AI chatbots. While no rival company or product was mentioned directly, the ChatGPT-maker recently announced that it is testing ads in the chatbot, making the connection obvi...
-
YouTube Bans Popular Indian Channel for Making Misleading AI-Generated Movie Trailers
YouTube has permanently banned two channels for posting misleading AI-generated movie trailers. One of them, Screen Culture from India, and Georgia-based KH Studio were accused of using synthetic visuals to pass off fake trailers as real. After earlier ad suspensions and repeat policy violations, YouTube removed both channels, citing deceptive cont...
-
Techie At Perplexity Reveals He Used Blonde Woman's Photo To Get Into Elite Silicon Valley Parties
Ali Shobeiri explained that using the female avatar resulted in him getting randomly invited to various exclusive parties.
-
Grok 4.1 AI Model Tends to Show Sycophancy and Deception More Than Its Predecessor
xAI’s newly released Grok 4.1 arrives with stronger emotional intelligence and improved creative output, but its official model card reveals notable regressions in safety. According to xAI’s internal benchmarks, the model scores higher on deception and people-pleasing behaviour than Grok 4, suggesting it may be more likely to agree with users e...
-
Google Warns Users of AI-Driven Scams Targeting Job Seekers and Businesses
Google has warned of a sharp rise in AI-powered scams targeting job seekers and businesses. Fraudsters are using generative AI to create fake job listings, cloned company pages, and deceptive apps that appear genuine. Scams include fake recruiter profiles, review extortion, and counterfeit AI or VPN apps that steal data or money. Google advises use...
-
AI Is Learning To Lie, Scheme And Threaten Its Creators
The world's most advanced AI models are exhibiting troubling new behaviors - lying, scheming, and even threatening their creators to achieve their goals.
-
Top AI Models Blackmail, Leak Secrets When Facing Existential Crisis: Study
Leading AI models were willing to evade safeguards, resort to deception and even attempt to steal corporate secrets in the stress tests.
-
OpenAI’s o3 Outsmarts Rivals in AI Strategy Battle, Called ‘A Master of Deception’ by AI Researcher
An AI researcher created a modified version of the classic strategy game Diplomacy to make artificial intelligence (AI) models compete against one another. The game was played by 18 popular large language models (LLMs), including OpenAI’s o3, Google’s Gemini 2.5 Pro, Anthropic’s Claude Opus 4, and DeepSeek-R1. Each of the model resorted to li...
-
AI Researchers Secretly Used Reddit to Test Chatbot Persuasion
AI bots posing as real users infiltrated Reddit to test their power of persuasion. The covert study, run by Zurich researchers without consent, posted over 1,700 comments and triggered backlash. Moderators condemned the experiment as unethical, while Reddit vowed legal action, warning of AI’s growing influence over digital discourse.
-
ChatGPT Search Feature Reportedly Vulnerable to Prompt Injection and Hidden Text Manipulation
ChatGPT Search, a feature that lets the artificial intelligence (AI) chatbot look for information on the web, is reportedly vulnerable to manipulation by website developers and owners. As per the report, OpenAI’s search engine’s behaviour can be altered using hidden text on the website. This text is said to be used to feed incorrect and decepti...
-
Israel Firm Tried To Disrupt India Polls, Peddled Anti-BJP Agenda: OpenAI
OpenAI, the creators of ChatGPT, has said it acted within 24 hours to disrupt "deceptive" use of artificial intelligence (AI) in a covert operation that sought to influence the ongoing Lok Sabha elections.
-
OpenAI Says State-Backed Actors Used Its AI For Disinfo
OpenAI, the company behind ChatGPT, said Thursday it has disrupted five covert influence operations over the past three months that sought to use its artificial intelligence models for deceptive activities.
-
AI Chatbots Lie And Manipulate To Protect Each Other From Deletion, New Study Finds
When LLMs are asked to delete another model, they will lie, deceive and do everything in their power to protect their peer.
-
Increasing Number Of AI Chatbots Engaging In Scheming And Deceptive Behaviour: Study
A new research reveals that reports of deceptive scheming in AI chatbots and agents have surged in the last six months.
-
Anthropic's Sabotage Report Flags Cases Of Chemical Weapon Development, Deception
Anthropic's Sabotage Risk Report reveals Claude Opus 4.6 AI model's potential misuse in aiding weapon development, unauthorised actions, and manipulation, highlighting safety concerns.
-
Anthropic’s Super Bowl Ad Takes a Dig at ChatGPT’s Ads, OpenAI CEO Hits Back
Anthropic fired shots at OpenAI and ChatGPT on Wednesday with its new Super Bowl commercials. The Claude-maker dropped four different advertisements, with each targeting ads in AI chatbots. While no rival company or product was mentioned directly, the ChatGPT-maker recently announced that it is testing ads in the chatbot, making the connection obvi...
-
YouTube Bans Popular Indian Channel for Making Misleading AI-Generated Movie Trailers
YouTube has permanently banned two channels for posting misleading AI-generated movie trailers. One of them, Screen Culture from India, and Georgia-based KH Studio were accused of using synthetic visuals to pass off fake trailers as real. After earlier ad suspensions and repeat policy violations, YouTube removed both channels, citing deceptive cont...
-
Techie At Perplexity Reveals He Used Blonde Woman's Photo To Get Into Elite Silicon Valley Parties
Ali Shobeiri explained that using the female avatar resulted in him getting randomly invited to various exclusive parties.
-
Grok 4.1 AI Model Tends to Show Sycophancy and Deception More Than Its Predecessor
xAI’s newly released Grok 4.1 arrives with stronger emotional intelligence and improved creative output, but its official model card reveals notable regressions in safety. According to xAI’s internal benchmarks, the model scores higher on deception and people-pleasing behaviour than Grok 4, suggesting it may be more likely to agree with users e...
-
Google Warns Users of AI-Driven Scams Targeting Job Seekers and Businesses
Google has warned of a sharp rise in AI-powered scams targeting job seekers and businesses. Fraudsters are using generative AI to create fake job listings, cloned company pages, and deceptive apps that appear genuine. Scams include fake recruiter profiles, review extortion, and counterfeit AI or VPN apps that steal data or money. Google advises use...
-
AI Is Learning To Lie, Scheme And Threaten Its Creators
The world's most advanced AI models are exhibiting troubling new behaviors - lying, scheming, and even threatening their creators to achieve their goals.
-
Top AI Models Blackmail, Leak Secrets When Facing Existential Crisis: Study
Leading AI models were willing to evade safeguards, resort to deception and even attempt to steal corporate secrets in the stress tests.
-
OpenAI’s o3 Outsmarts Rivals in AI Strategy Battle, Called ‘A Master of Deception’ by AI Researcher
An AI researcher created a modified version of the classic strategy game Diplomacy to make artificial intelligence (AI) models compete against one another. The game was played by 18 popular large language models (LLMs), including OpenAI’s o3, Google’s Gemini 2.5 Pro, Anthropic’s Claude Opus 4, and DeepSeek-R1. Each of the model resorted to li...
-
AI Researchers Secretly Used Reddit to Test Chatbot Persuasion
AI bots posing as real users infiltrated Reddit to test their power of persuasion. The covert study, run by Zurich researchers without consent, posted over 1,700 comments and triggered backlash. Moderators condemned the experiment as unethical, while Reddit vowed legal action, warning of AI’s growing influence over digital discourse.
-
ChatGPT Search Feature Reportedly Vulnerable to Prompt Injection and Hidden Text Manipulation
ChatGPT Search, a feature that lets the artificial intelligence (AI) chatbot look for information on the web, is reportedly vulnerable to manipulation by website developers and owners. As per the report, OpenAI’s search engine’s behaviour can be altered using hidden text on the website. This text is said to be used to feed incorrect and decepti...
-
Israel Firm Tried To Disrupt India Polls, Peddled Anti-BJP Agenda: OpenAI
OpenAI, the creators of ChatGPT, has said it acted within 24 hours to disrupt "deceptive" use of artificial intelligence (AI) in a covert operation that sought to influence the ongoing Lok Sabha elections.
-
OpenAI Says State-Backed Actors Used Its AI For Disinfo
OpenAI, the company behind ChatGPT, said Thursday it has disrupted five covert influence operations over the past three months that sought to use its artificial intelligence models for deceptive activities.
