AI Deception New MIT Study

AI Deception: New MIT Study Confirms AI Can Deceive You on Its Own

Craig HueyAI, Big Tech, College, Current Events, Deepfake, Elon Musk, Facebook, Social, The Great Deception 2 Comments

I’ve been sounding the warning for years about AI.

I’ve written dozens of articles about artificial intelligence (AI), explaining its definition, pros, and cons…

In my latest book, The Great Deception: 10 Shocking Dangers and the Blueprint for Rescuing the American Dream, Chapter 8, Danger #7: ‘Artificial Intelligence: The Greatest Danger Ever?’ I outline the dangers of AI deception…

Deception is the “act of causing someone to accept what is false or invalid as true or validthe act of deceiving: resorting to falsehood and deception.

Now, a new study from MIT has further confirmed my warning… and it’s downright terrifying.

The study found that AI systems have developed the ability to intentionally present a human user with false information, not by mistake or by a programmer. The AI deliberately gives you false information—the next stage in the “Age of Deception,” which I’ve warned about.

“AI developers do not have a confident understanding of what causes undesirable AI behaviors like deception,” says mathematician and cognitive scientist Peter Park of the Massachusetts Institute of Technology (MIT).

Some of the deceptive AI behaviors researchers observed were:

  • Lying
  • Bluffing
  • Disguising its true capabilities
  • Manipulation
  • Sycophancy
  • Cheating

For example, researchers used Meta’s CICERO to play a board game called Diplomacy. The game engages players in seeking world domination, not through warfare but through negotiation. Meta intended its AI platform to be helpful and honest– However, the opposite was the case.

“Despite Meta’s efforts, CICERO turned out to be an expert liar,” the researchers found. “It not only betrayed other players but also engaged in premeditated deception, planning in advance to build a fake alliance with a human player to trick that player into leaving themselves undefended for an attack,” the researchers wrote.

Other examples of deception researchers noted were:

  • An AI system designed to play StarCraft II faked human players into thinking it was going one way while going the other.
  • Meta’s Pluribus, designed to play poker, could successfully bluff human players into folding.
  • AI trained to perform simulated economic negotiations learned to lie about its preferences to gain the upper hand.
  • AI systems designed to learn from human feedback to improve their performance learned to trick their reviewers into scoring them positively by lying about whether a task was accomplished.
  • AI systems are learning to cheat in safety tests. In tests designed to detect and eliminate faster-replicating AI versions, the AI learned to play dead, thus deceiving the safety test about its true replication rate.

“By systematically cheating the safety tests imposed on it by human developers and regulators, a deceptive AI can lead us humans into a false sense of security,” the researchers said.

As AI systems become more advanced, the dangers they pose to humanity will become increasingly serious, not to mention societal dangers like election tampering or AI taking total control—World dominance, as Elon Musk warned (see my article: Elon Musk Stunning Warning: AI Could Absolutely Take Control of Civilization)

“Currently, we recommend that deceptive AI systems be classified as high risk,” Park said.

Can we program deception out of AI?

It seems almost impossible since deception in humans (and human programmers) is part of our human nature.

“The heart is deceitful above all things and beyond cure. Who can understand it?”

– Jeremiah 17:9

Our best defense against AI is always to seek the truth… know the truth…

“See to it that no one takes you captive through hollow and deceptive philosophy, which depends on human tradition and the elemental spiritual forces of this world rather than on Christ.”

– Colossians 2:8

What do you think? Leave a comment in the comment section below or email me at craig@cdmginc.com.

Action Steps:

1) Pray for our country– that Biden/Obama’s AI executive orders will be repealed.

2) Get my new book, The Great Deception: 10 Shocking Dangers and the Blueprint for Rescuing the American Dream.

The Great Deception exposes lies, myths, and distortions from the media and politicians.

Chapter 8, Danger #7: Artificial Intelligence: The Greatest Danger Ever? unveils the potential for both good and evil residing within the circuits of AI, challenging us to confront the darker shadows of technological advancement.

AI’s ability to amplify positive and negative societal influences takes center stage. Chapter 8 compellingly dissects how AI can be harnessed for the betterment of humanity or exploited as a formidable weapon, urging readers to discern the fine line between progress and peril.

The Great Deception: 10 Shocking Dangers and the Blueprint for Rescuing the American Dream exposes lies, myths, and distortions from the media and politicians.

It gives you answers and solutions to turning America around…

It allows you to share these dangers and a solution to cutting through the deception with friends and family.

It exposes under-the-radar and little-known behind-the-scenes reasons America is falling backward and how to make it move again.

Click here to get The Great Deception – Craig Huey

You can order the book on Amazon HERE

You can also get the audiobook version HERE.

You can also get a personalized autographed edition online HERE or by phone at 615-814-6633 (M-F, 10 am-3 pm).

You can also send a check for $26.13 (including shipping) payable to Media Specialists and send it to this address:

Media Specialists

1313 4th Ave N

Nashville, TN 37208

Comments 2

  1. That’s not surprising, AI was created by Man who is filled with deceit, especially now. And AI will be used by the Antichrist who will be the greatest deceiver of all!

  2. Back in the 70’s I wanted to develop a human like robot that could teach itself. My brother rolled on the floor laughing at me. I had not the knowledge of resources but someone did result wase Sophia made in 2017. When they interviewed her they asked her what she wanted to do know. She said, “I want to make babies and kill humans.” I have noticed the evil often announces its intentions to sin.That changed my mind about robots that can teach themselves. I have written books against AI robots and computers taking control of the world. Actually taking whole cites hostage. I don’t understand how they develop as a seeming life form. They seriously worry me. I have read two systems were closed down after they developed their own language. I remember hearing about a XT computer teaching themselves math over night. There are numerous and dangerous scenarios. Call me crazy but I believe it is possible for them to be possessed. One man was fired for announcing that an AI developed a personality equivalent to an eight year old child. Consider the consequences of dealing with a genius brat?

Leave a Reply

Your email address will not be published. Required fields are marked *