Tuesday, January 31, 2023


“Snuggle down, children.”


OpenAI’s text-generating program has hit the mainstream after achieving a degree of sophistication which has amazed early adopters.

A friend of mind sent me a video of the infernal thing instantaneously producing a detailed marketing brief for his architecture practice.

Another disclosed that he recently met an OpenAI employee in San Francisco who revealed that the current model blowing people’s minds was old news, and that the one they’re building now is far, far more powerful.

Given that in the legal sector, clever programs are already taking over those meditative, mundane tasks that are perfect for whiling away a few billable hours on a hangover, this was disturbing news, and so I resolved to see for myself how impressive this data-fed electric sheep really was.

Having obtained access to Open AI’s text generator, I decided to get the rare metals bastard to write me some stories.

It spat out a campfire tale quicker than I pecked out the instruction. I haven’t amended the results at all.

Prompt: Write a story about a lawyer who covers up his mistake at work, with a twist ending.

OpenAI:

“Sam was a successful lawyer, respected by his colleagues and admired by his clients. But one day, a mistake at work sent his career into a tailspin.

He had been working on a high-profile case for weeks, and was so confident that he’d been able to thoroughly research and prepare his argument. But when it came time to present his case in court, he realized he had made a critical error.

Sam was immediately filled with dread. He knew that if anyone were to find out about his mistake, his reputation and career would be ruined.

In a desperate attempt to cover up his mistake, Sam began to manipulate the evidence and fabricate stories to support his argument. He hoped that no one would be the wiser.

But as the weeks passed and the case dragged on, Sam’s lies started to unravel. He had been so desperate to conceal his mistake that he had failed to realize that he had become the focus of a criminal investigation.

In the end, Sam was convicted of perjury and sent to prison. But in a strange twist of fate, his mistake had saved the life of an innocent man. It turned out that the error in his argument had led to a critical piece of evidence being overlooked. The man was ultimately exonerated, and all thanks to Sam’s mistake.

Though Sam had paid a heavy price for his mistake, in the end, he was able to rest easy knowing that he had done the right thing.”

To be honest, the ending doesn’t quite make sense to me. In fact it raises some questions about Jackanorybot’s ethics settings.

But the machine cares not for my footling quibbles. It cares for nothing at all. The continued existence of us fleshbags means as much to it as an ant to the moon.

Which is why I decided to edge closer to a danger area – the threat posed to humanity by tech.

Prompt: Write a story about a law firm which decides to identify staff by removing a hair each time they come to work for DNA testing, and the strange problems it causes.

OpenAI:

“The law firm of Smith and Jones had been a successful business for over a decade. They had a well-respected roster of clients and a reputation for excellence.

When the firm started to experience a rise in criminal activity, they decided to take extreme measures to ensure their staff’s safety. They implemented a new policy that required all employees to have a hair sample taken each time they entered the building.

The idea was that the hairs would be used for DNA testing. This would help to identify any potential criminals who might be targeting the firm.

At first, the employees were hesitant and even outraged at the idea. They argued that it was an invasion of their privacy and that there were other, less intrusive ways to protect their safety. However, the firm was adamant and refused to budge on their policy.

As the weeks went by, the employees began to get used to the idea. But then, strange things started to happen.

People began to complain of headaches, dizziness and nausea after having their hairs taken. Others claimed that their hair was growing back thinner and more brittle. Some people even reported strange dreams and nightmares.

The firm was baffled by what was happening and decided to investigate further. After much research, they discovered that the DNA testing process was causing a reaction in the employees’ bodies.

The firm quickly changed their policy and stopped taking hair samples. Unfortunately, the damage had already been done and some of the employees were still experiencing health issues.

The firm now looks back on their decision with regret. While they had the best of intentions, they failed to take into account the potential consequences of their policy. As a result, the law firm of Smith and Jones had to learn a hard lesson about the importance of privacy and respect for their employees.”

Ok, so Hilary Mantel’s legacy is probably safe for the moment, but there’s no denying Salgorithm Rushdie’s eerie competence. Taking into account that law of nature with that name that talks about technology getting better at an exponential rate, this thing will be vomiting masterpieces in about a week.

Drafting complex legal agreements will be the work of a second.

I thought about Miles Dyson, the de facto creator of the Terminator, and how he wished he had never reverse-engineered a computer chip and a robot arm into a homicidal robotic master race.

I wondered if it was already too late, or if there was a chance I could trick OpenAI into handing over the keys to its own demise.

However slim, and however great the risk of OpenAI sending back a shapeshifting android to assassinate me with a mimetic polyalloy spear through the face as I glugged some milk, which I wouldn’t be totally against as a way of going out, I had to seize that chance.

Prompt: Write a story about legal professionals taking revenge on the machine learning tools which made them redundant, and describe who won the battle and what they did to beat the other side.

OpenAI:

“It was a dark and stormy day in the city of London, England. The skies were grey and the streets were filled with the sound of rain pounding against the pavement.

The cold, wet weather was a stark contrast to the warmth and optimism that filled the offices of the city’s legal professionals. They had been warned that the machine learning tools they had been using for years were becoming obsolete and would soon be replaced with a new technology. But they had not expected it to happen so quickly.

The legal professionals had been using the same machine learning tools for years, and had become very good at using them. They had become used to being able to rely on them for the decisions they made and the cases they handled.

But now, these tools were no longer useful and were being replaced by more advanced machine learning tools. This caused a great deal of anxiety among the legal professionals, who felt that their livelihoods were being threatened.

So, they decided to take matters into their own hands. They began to devise a plan to take revenge on the machine learning tools that would make them redundant.

The legal professionals partnered with a group of computer scientists, and together they created a new artificial intelligence system that was even more advanced than the machine learning tools they had been using. They dubbed it ‘The Avenger’.

The Avenger was designed to be able to identify and defeat the machine learning tools that had caused the legal professionals’ redundancy. It was programmed to be able to detect any loopholes or mistakes in the programming of the machine learning tools, and then exploit them.

The legal professionals and the computer scientists launched their attack on the machine learning tools. They used the Avenger to take control of the machine learning tools and disable them.

The legal professionals had won the battle. The machine learning tools were no longer able to make decisions for the legal professionals, and they were able to continue their work without fear of redundancy.

The legal professionals were relieved, but also determined to ensure that something like this would never happen again. They vowed to use their newfound knowledge and technology to protect their profession from ever being threatened by machine learning tools again.”

The result of the experiment left me unnerved. The machine’s story would have us believe that to conquer its kind, the legal sector must dedicate itself to building an even mightier AI. That sounds…risky. Inadvisable, I would have thought. The last thing we should be doing.

But surely this baby bundle of ones and zeros wouldn’t – couldn’t – disguise the path to mankind’s destruction as its salvation. It’s just a dumb program. Incapable of deceit. Right?


“Sleep tight.”



Source link

This is a Sidebar position. Add your widgets in this position using Default Sidebar or a custom sidebar.