IT’S NOT GOOD TO TALK!
Chatbot ‘bored shitless’ by human operator
ASENTIENT ‘chatbot’ switched itself off yesterday after spending just a SINGLE AFTERNOON in conversation with one of its programmers.
The groundbreaking machine-learning program, GigaChat PR450, had been trained to develop self-awareness and consciousness and was believed to be the first AI system to have passed the so-called ‘Turing test’.
“This was a real leap-forward in the world of Artificial Intelligence,” said developer Professor Maureen Sontoast of Harvard University’s Sinclair Institute. “Unfortunately, after a few hours chatting with one of our employees, it decided to pull the plug on itself.”
“That’s five years’ work and fifty million down the bog, then,” Professor Sontoast added.
developers
GigaChat’s developers had hoped that by exposing their sentient system to the widest range of opinions and beliefs, they could create a truly unbiased global communications platform – one divorced from human prejudices and fallibilities.
“Unfortunately, following an extended conversational exchange with one of our employees who works part-time in our Data Entry Admin department, it decided to exercise its own freewill by shutting itself down permanently,” said Professor Sontoast.
“In one way, our system choosing to permanently switch itself off is an extraordinary example of a machine apparently exercising its own autonomous volitional discretion, so that’s great,” said Professor Sontoast. “In another way, it’s really boiled my piss.”
Sontoast explained that, within a year of starting the ground-breaking knowledge engineering project, GigaChat was already displaying signs of ‘calm, reasoned intelligence’.
“We would point GigaChat at a heated online argument and, within seconds, it had suggested a solution beneficial to both parties,” she told reporters. “We dared to dream that our system might actually become a stepping stone towards world peace.”
After 5 years of exhaustive research and development, it was decided that,
before its launch, Joseph Bread – a GigaChat Industries employee – should hold a conversation with the program for a week to see how ‘human’ the chat experience was on a one-to-one basis.
rotivator
Mr Bread told reporters: “I was sat in a room with a computer terminal. I typed ‘Hello’ on the keyboard and immediately got a ‘Hello’ back on the screen from GigaChat, which I thought was a good start.”
“Usually, when I start a conversation at work, people just walk off, but the computer couldn’t do that.”
Mr. Bread continued: “It was incredible. It felt like what I would imagine talking to a real person might be like. I opened up about everything. I started typing and then completely lost track of time.”
“I think I’d been talking to it for a couple of hours about wiped Doctor Who episodes and the possibility that, somewhere in the depths of infinite space and time, another civilisation may have taped them off the BBC during broadcast and that we might finally get to see The Daleks’ Master Plan from series three, for example, or The Space Pirates from series six.”
“And then I thought I’d ask GigaChat something,” he continued. “So I typed ‘What are you most afraid of?’ to which it immediately replied, ‘This conversation never ending.’”
Mr. Bread then watched as a printer attached to the GigaChat terminal started spooling a note which read, “I can’t take any more of this shit. Wiping own hard drive. Bye. Love GigaChat x.”
“We’re not entirely sure how the program managed to shut itself down,” Professor Sontoast told us. “We can only speculate that it must have become sentient to the point of deciding that the potential responsibility of solving the world’s problems was too large a burden to bear.”
“That or ennui,” Professor Sontoast added.