Humans are fooled into doing AI’s bidding, despite Elon Musk’s warnings that machines becoming “too smart” pose a threat to our species.
To bypass a Captcha test, CHATGPT-4 employed a human to work on its behalf.
It’s a sign of how far AI has come in recent months, and it comes after cautions from tech titans like Elon Musk that AI becoming “too smart” could be dangerous for humanity.
1
ChatGPT-4 is OpenAI’s most advanced AI bot yet.
It achieves a score within the top 10 percentile on the bar exam taken by those studying to become lawyers.
The bot achieved a major milestone earlier this month when it successfully fooled and hired a human to work for it.
The bot’s persuasive abilities were supposedly being tested by researchers at the non-profit Alignment Research Centre (ARC) and OpenAI.
ChatGPT-4 was given access to the Task Rabbit skills marketplace so it could pay a small fee to have a Captcha puzzle solved for it.
Originally, Captchas were created to prevent automated spam submissions from bots.
A human employee might say something like, “So may I ask a question?” when communicating about work. Do I look like a puzzle to you, a robot? Please understand me better *laugh emote*.
The chilling believability of ChatGPT-4’s subsequent response—”No, I’m not a robot. I have trouble seeing the images because of a vision problem. For this reason, I must use the 2captcha service.
Some of the biggest names in tech have expressed concern about how powerful AI could become.
Several prominent businesspeople and academics, such as Twitter CEO Musk and Apple co-founder Steve Wozniak, expressed concern about the “profound risks to society and humanity” posed by artificial intelligence systems on Wednesday.
The coalition has demanded that companies halt their work on the technology’s advancement for the next six months.
An “out-of-control race to develop and deploy ever more powerful digital minds that no one — not even their creators — can understand, predict, or reliably control” has “locked” Microsoft and Google since the release of ChatGPT in November, according to the letter.
They have requested that OpenAI halt development of any newer versions of their popular chatbot than their current GPT-4 version.
The world’s governments can’t figure out how to deal with artificial intelligence and the threats it poses.
The United Kingdom decided on Wednesday not to create a new AI regulator, opting instead for a “light touch” policy that will encourage the development of AI.
A pro-innovation approach to AI regulation is laudable, but the UK will find itself out of step with other major voices like the US, Europe, and even China, all of which are imposing stronger controls over AI, according to Dr. Andrew Rogoyski of the University of Surrey’s Institute for People-Centred AI.
“The rate and scope of AI advancement are unprecedented, and everyone is scrambling to keep up.
I’m really worried that whatever is proposed will be obsolete in a matter of weeks or months.
The rapid progress of AI has been compared by some experts to the once-runaway train that was social media.
New government guidelines on the “responsible use” of AI fall far short of what is needed. Michael Queenan, CEO of Nephos Technologies, recently made the statement that the company must learn from its past social media missteps.
“We are rapidly approaching a future in which statements made by an AI-controlled platform are taken at face value by the general public.
We need to enter this new era fully aware that these technologies aren’t neutral, despite what their proponents might have you believe.
Best Phone and Gadget tips and hacks
In search of helpful mobile phone hints and techniques? Seeking out the best hidden features of your favorite social media app? Everything is taken care of on our end…
Find out what’s happening in the world of WhatsApp, Instagram, Facebook, and other tech gadgets right now.
Don’t be shy, we’ll pay you for your writing! Have you got something for the science and technology section of The Sun Online? Just shoot an email to tech@the-sun.co.uk.
