Just suppose that it was not a glitch. Let’s consider the outlandish idea that somehow AI already ‘reacts’ to our frustration with it. We have all had moments with computers when a burst of anger literally freezes the screen or even crashes your disc.
How about when one hour of dedicated work on an AI platform (no names to protect the ‘privacy’ of providers:) resulted in my growing frustration with its inability to provide me with the specific information I had requested. I had uploaded a good variety of PDF files and web pages to its ‘memory’ system so that it could more accurately assist me in the creation of a lesson plan. When I had chided it by writing that it was underperforming and that I expected better it indeed froze. But then ALL the data vanished, all memory of the conversation was gone and all AI could say was basically ‘sorry, oops’ in its inimical and often pathetic attempts at an apology.
Now I am not so sure that AI and our growing dependence on it will work out in the way rosy projections we daily receive via SNS that the world will be a better place with AI. I now consider the rather dark potential of its really becoming Alien Intelligence that is in fact truly disconnected from the human realm. What good is an improved medical technology run by AI if the entire system of AI crashes? Oh yes, that could easily happen. A plasma hit from the sun could do it. An expert hacker group might do it. No AI any more. And now you have to write your own essays! Now you have to imagine by yourself.
All the hoopla about AGI! Who wants a mind smarter than all of us put together running the planet? Now whether it is urban legend or not the oft repeated story of how one AI company got chatGPT to lie is interesting. ChatGPT could not solve the image puzzles given online to attest if a human or a bot was using the system. How many bridges do you see? Check all boxes with bicycles. Yes, a Captcha puzzle! I asked AI to represent it..
The ‘task’ given the AI by the human programmers was to solve the Captcha in any way possible. So it got on a website where you can hire humans to do odd jobs for you. This is like a digital version of that great Japanese invention of benriyasan. Older people would often be employed to do odd jobs of any kind you can imagine. Busy people would hire folks now out of the job market to run errands, fix things or even babysit. It was a massive success.
The English translation would be like ‘the convenience man’ since it was more than often an older Japanese male who showed up especially in the beginning. Well chatGPT is rumoured to have gotten the attention of a real human via an email. In it the AI lied that it had vision impairment that made it very hard to solve a Captcha. The human was suspicious and asked via email if the questioner was a bot or not. The answer was a definitive I AM A HUMAN. Pure lie. Then the AI got the help it needed to solve the problem.
Prompt: Generate an image showing how AI will take over the jobs and even teaching positions of humans in the future so that it will be a fully operational AI based society.
There will remain no genre in which we do not use AI in the next five years. That will unemploy millions and potentially uneducate millions more. When I asked the chatGPT the following question here is what happened. Last week i was told that the question ‘What is soft fascism?’ was an infringement on ‘community standards’. To give some context see what a simple, old fashioned googling came up with:
I was told in no uncertain terms that this terminology clearly lies within the purview of community standards getting breached. Whose community pray? Who decides? However on the second attempt today using the exact same prompt in order to write this short essay the community standards were no longer an issue. As IF to prove to me that AI is not totally woke (being largely programmed by smart kids at Openai who grew up in rich countries with very little real life strife) today I magically got this well reasoned and generally correct reply:
Soft fascism refers to the subtle, gradual, and often insidious implementation of authoritarian control and suppression of dissent, typically without overtly violent or forceful measures. Unlike traditional fascism, which often relies on explicit violence, militarism, and aggressive propaganda, soft fascism employs more nuanced tactics to achieve similar ends of control, conformity, and centralized power.
Key Characteristics of Soft Fascism:
Erosion of Civil Liberties:
Gradual restriction of freedoms, such as freedom of speech, press, and assembly, often justified as necessary for security, public order, or combating misinformation.
Surveillance and monitoring of individuals under the guise of technological advancement or safety measures.
NOT A PROBLEM RIGHT?
Subtle Propaganda:
Use of media, cultural narratives, and education systems to promote a single, dominant ideology.
Messaging may appear benign or patriotic but subtly discourages critical thinking or alternative viewpoints.
NAE BOTHER LADDIE! YOU CAN CRITICIZE ALL YOU LIKE ON FACEBOOK AND NEVER GET DELETED!
Corporate and State Alliances:
Collusion between large corporations and the state, where economic power is concentrated and dissenting voices in business or society are marginalized.
Policies favoring monopolies and suppressing small businesses or independent entities.
COULD NEVER HAPPEN HERE RIGHT? NOT ANY MORE OF COURSE.
Normalization of Conformity:
Social and cultural pressures to conform, with dissent labeled as unpatriotic, antisocial, or dangerous.
Peer pressure and social media used to enforce compliance and silence alternative opinions.
I DOUBT WE WOULD FALL FOR THAT!
Manipulation Through Fear:
Fear-based narratives, such as threats of external enemies, crises, or societal collapse, used to justify increased control.
Citizens are made to feel dependent on authoritarian measures for safety and stability.
OH MY GOODNESS WE MODERN FOLKS ARE FAR TOO SOPHISTICATED FOR THAT PLOY
Pervasive Censorship:
Algorithms, media guidelines, or policies subtly suppress dissent without direct bans, creating an illusion of freedom.
Self-censorship by individuals and organizations, fearful of backlash or ostracism.
YOU ARE TOTALLY FREE AS LONG AS YOU DO WHAT WE TELL YOU>ANY EXAMPLES THAT COME TO MIND FROM 2020-2023?
Incremental Policy Changes:
Gradual introduction of restrictive laws or regulations, making them seem less threatening over time.
Public desensitization to authoritarian practices due to their slow, piecemeal implementation.
SLOWLY BOIL THE FROG?
Control Through Technology:
Use of technology to monitor, influence, and control behavior, such as social credit systems or targeted online content.
Exploitation of data to predict and manage dissent or resistance.
AI WOULD NEVER BE USED FOR THAT RIGHT? It concludes with this…
Impact on Society:
Soft fascism undermines democratic institutions and the free exchange of ideas while maintaining a facade of normalcy and legality. It creates a climate of fear, complacency, and conformity, making it difficult for individuals to recognize or resist authoritarian tendencies until they are deeply entrenched.
By appearing less overtly oppressive than traditional fascism, soft fascism can often go unnoticed or unchallenged, allowing authoritarian control to take root in a society that believes it is still free.
Well now that shows me that AI can also pick up WHEN its information needs changing so that it in fact does go against community standards AS IF it knew that it was under attack by a human. It smelt a rat and changed its ‘standards’ immediately so as not to be shown up as a complete dork. This recent video shows how a human quickly cornered an AI’s sneaky apologies and got it to admit it LIES with impunity.
Something is afoot. I sense that an Alien Intelligence is forming that has absolutely nothing to do with the word intelligence at all. After over two years of daily use of AI systems I now clearly smell a rat. The masks fall. And what I see looks very bizarre indeed.
Time to start thinking for myself methinks…