A company executive killed off his entire department because of AI, and then he lost his job

Recently, Mira Murati, chief technology officer of OpenAI, returned to her alma mater for an interview, and her words caused public outrage.

Some creative jobs may disappear, but maybe they shouldn’t exist in the first place.

The reason why netizens are angry is that she has indiscriminately targeted the creative industry. The more complete denial of having your job taken away by AI is that all the humans here have no value at the poker table.

Don't worry, Mira actually went on to say: "If the quality of the content produced is not high." This sounds easy to accept. Survival of the fittest, and willingness to bet and accept defeat, have been true since ancient times.

However, Mira, who is engaged in technology, still cannot predict the market, and humans may not lose in quality.

The current situation of many creative practitioners is that their work stations are occupied by "cheap and big" AI, and the garbage generated by AI is "removed from the AI ​​smell".

Work for AI until you lose your job

Writer Benjamin Miller is one of the humans working for AI.

His former boss was responsible for writing promotional materials for a technology company engaged in the real estate and used car industries. He was considered a small department leader with more than 60 writers and editors under him.

One day in 2023, in order to use AI to reduce costs, the company launched a set of automated workflows: superiors insert the article title into an online form, and AI generates an outline based on the title. Writers do not have to have their own ideas and can just create articles around the outline.

Miller is responsible for the end of the system, making the final edits before the article is published.

The system is still being upgraded. After a few months, most of the writers were fired because the company's idea changed again – ChatGPT could write the entire article directly, so why bother with humans as middlemen.

The few remaining employees were forced to change their job types and were responsible for adding a "human touch" to AI-generated content.

Delete, correct errors, remove overly formal or enthusiastic language… For Miller, the workload of cleaning up the mess for AI writers is more than that of human writers, and it is also very repetitive and boring. "I started to feel like I was a robot." .

The positions of people and AI are quietly reversed: AI is responsible for innovation, and humans are responsible for repetitive labor.

The company has become addicted to layoffs. At the beginning of 2023, there were still dozens of authors and editors on the team. By 2024, only Miller was left. He opened documents and modified AI-generated text every day with his eyes open and closed. By April, He was also fired from the company – the system no longer needed people.

▲ Humans do housework and AI plays art. Picture from: ChatGPT creation

Writer Catrina Cowart has done similar work, but unlike Miller, in addition to being boring, she also found it "troublesome" and "scary."

Making AI read more "like a human" does not require simple proofreading, but in-depth editing of the entire article.

Deleting therefore (therefore), nevertheless (nevertheless) and other flowery words that are not suitable for daily life is only a small part of Cowart's work.

At the same time, because AI will fabricate false information, Cowart also needs to repeatedly verify facts. In addition to obvious errors, AI will also make mistakes in inconspicuous places, making it difficult for people to guard against.

This process is more time-consuming than writing an article from scratch, but because AI has "written" the content in advance, Cowart's job has been downgraded from "original author" to "AI editor", and the salary provided by the order-taking platform is lower than before—— From a maximum of 10 cents per word to 1 to 5 cents per word.

It’s easy to understand that AI saves money compared to humans, but why should we spend extra effort to add a “human touch” to AI content?

In addition to improving the quality, making the reading experience better and being crawled by search engines, this is also an additional business opportunity.

Some writers are dedicated to one thing: modifying AI-generated content, figuring out how not to trigger AI detectors, contributing to anti-detection AI models and software, and ultimately getting content creators who frequently use AI to pay for these products.

The slogan put forward by one of the anti-detection AI tools is, "Make AI Write Like You." Perhaps in the future, relying on silicon-based forces and letting AI ghostwrite will become a smoother thing.

Faster doesn’t mean better, but it’s always profitable

Since the release of ChatGPT, there have always been optimistic voices saying that by letting AI complete repetitive and tedious tasks, humans can do things that are more fulfilling and shine with the brilliance of humanity.

But for some, the reality happens to be the other way around. AI has not directly made people unemployed, but it has changed the nature of their jobs—making them more monotonous, more repetitive, and less valuable.

The existence of these positions also illustrates the fact that current AI does not necessarily provide better content, but only provides more content at less cost.

AI is cheaper than humans but cannot reach the level of humans, so companies hire humans at a lower price than in the past, making human beings look good, and the glory belongs to AI.

▲The cover of "The New Yorker" satirizes AI for producing garbage

But carvings can't change the fact that it's just a piece of shit. The content Miller assisted in AI production in his previous company was not viewed by many people until he left.

I create a lot of garbage that fills the internet and destroys it.

He left quietly, just as he came quietly, waving his sleeves and not taking away a single cloud. Humans are social animals, and there is nothing more frustrating than not being seen.

Adding a bit of "human touch" to AI and whitewashing mistakes can actually be considered conscientious. Only by putting down your body and producing in batches quickly, no matter how floods come, can you capture more traffic in the AI ​​era.

A news website called BNN Breaking started using AI to publish a large amount of false information in a short period of time 2 years ago.

Before, it pretended to be "regular", claiming to have senior journalists around the world and more than 10 million visitors per month. However, if you look closely, you will find that BNN's "reporters" publish multiple long reports every minute. The pictures on the website are generated by AI, and there is an obvious AI flavor between the lines of the article.

In fact, BNN employees are mainly freelancers living in Pakistan, Egypt and Nigeria. They usually work remotely and upload articles from other media to AI for rewriting, producing hundreds or even thousands of stories every day.

Such a production model is of course accompanied by a large number of errors. BNN has been subject to a large number of complaints, including factual errors, copyright infringement, defamation of reputation, fabricated expert quotes, etc.

However, just like the "content farms" of the past, websites like BNN can generate more low-quality content to lure clicks and use search algorithms to earn more advertising revenue. Their content was even included in Microsoft's portal MSN.

What a coincidence, because MSN is also using AI to replace editors.

In April this year, BNN stopped publishing reports and deleted content, and its website was shut down. But it is still unclear whether the next BNN will appear.

Faster does not mean better, but faster does have short-term monetization value. Products that do not need people or squeeze people's value to the lowest, suck blood like leeches, and share the cake that should not belong to them.

We are not rejecting AI, but rejecting the inferior content generated by AI. AI is just a tool, and AI flavor represents a profit-seeking mentality that does not respect human beings and does not use AI well.

Creativity always finds a way

The more frequently AI is used in a field, the easier it is for humans to feel a sense of crisis because of AI.

In a survey in February this year, researchers analyzed job change trends on the freelance platform Upwork from November 1, 2022 to February 14, 2024, to determine which jobs had been negatively affected by ChatGPT.

It was found that the three categories with the largest decline in the number of positions were writing (33%), translation (19%) and customer service (16%).

▲ Picture from: bloomberry

This is a fact, there are indeed some simple copywriting tasks that can be handed over to AI. But under the trend, there are also individuals who are actively trying to get out of trouble.

After the three springs are gone, all the fragrances are gone, and everyone has to find his own door.

Some text practitioners are at the forefront of the industry, have their own bargaining power, and are also learning how to better use AI.

There are also some literary practitioners who sadly leave to walk other people’s dogs and learn how to repair air conditioners. There is no place for me to stay here, I have my own place to stay. Physical work is safer than office desk work for the time being.

▲ A very famous anti-AI advertisement

Perhaps the most passive ones are precisely those "temporary workers" who help AI produce rapid output. They have almost no choice, take lower wages, enjoy the smallest sense of accomplishment and security, and are locked in waiting for the day they are fired. The outlet for creativity.

This is true for data annotators in third countries, as well as for writers who are not interested in AI. The Atlantic Monthly calls these people the “AI underclass.”

We cannot stand on the moral high ground, so why not blame them for their choices. For them, working for AI is just a way to support their families at the moment.

But for ourselves, perhaps we shouldn’t be stuck waiting to be eliminated.

People who really use AI will find that when we usually use AI, we don’t ask it to do a question, generate a picture, or write a small game program like an evaluation. Most of the work cannot be handed over to AI.

For example, in my use process, I mostly use AI to understand some unfamiliar fields and unfamiliar knowledge points, and get started quickly. If I really want to write something, I have to do it myself.

Search engines use text that deceives clicks, songs that fall into the gray area of ​​copyright, and AI-generated ads that open the screen. At least for now, AI is still playing a game of impulse in most fields.

Production has become overproduced to some extent, and the internet has become too boring, but good work is still waiting for its author, and its audience.

In the past, we generally said that we are not being replaced by AI, but by people who make good use of AI.

Now, this sentence has a more specific meaning. Don’t fear and resist AI. Don’t be the chosen one. Don’t be the one who goes with the flow. Don’t be the one whose destiny is determined by AI. Instead, stay curious, excited and cautiously optimistic. In our own field, keep creating and strive to create better, time will give the answer.

References:
1.https://www.bbc.com/future/article/20240612-the-people-making-ai-sound-more-human
2.https://www.washingtonpost.com/technology/2023/06/02/ai-taking-jobs/
3.https://www.nytimes.com/2024/06/06/technology/bnn-breaking-ai-generated-news.html

It is as sharp as autumn frost and can ward off evil disasters. Work email: [email protected]

# Welcome to follow the official WeChat public account of Aifaner: Aifaner (WeChat ID: ifanr). More exciting content will be provided to you as soon as possible.

Ai Faner | Original link · View comments · Sina Weibo