Redazione RHC : 7 August 2025 07:06
The wave of layoffs in Silicon Valley continues in 2025.
Silicon Valley CEOs are shouting that artificial intelligence will bring a new revolution in efficiency and will also be the light of the future for reshaping productivity. However, when this technological dividend falls on the heads of ordinary workers, there is often only one seemingly valid reason for the layoffs.
The cost savings achieved through the use of artificial intelligence appear to have become an important element in corporate financial reporting, but for it to really work and be used stably, more and more companies must increase investments in subsequent maintenance, content review, security and compliance, and other aspects.
As a result, a new profession has quietly emerged: cleaning up AI-caused problems.
It’s mainly about rework, remediation, and customer support, fixing sites compromised by AI.
AI tools have invaded internal business processes and represent the most visible business trend of the last two years.
OpenAI, Google Gemini, Anthropic Claude.… Models emerge one after the other The other, and products/functions are increasingly focused on the workplace, the most productive scenario: writing texts, editing code, generating scripts for customer service, artificial intelligence is used as a magical tool for work, as if its use could reduce costs and increase efficiency overnight.
Anthropic CEO Dario Amodei said in an interview in May that in the next one to five years, half of entry-level jobs will be replaced by artificial intelligence and the unemployment rate in the United States could rise to 10-20%.
But a recent BBC report revealed the other side of this “use” of AI: many companies that have saved budgets thanks to AI are spending more money on rework and consequences. Sarah Skidd is a freelance copywriter working in the United States. In May of this year, she received an urgent assignment from a content agency: to modify the AI-based copywriting for all pages of a hotel website.
20 hours of work, at an hourly rate of $100, for a total of $2,000: the money that was “originally intended to be saved” was respent in the form of higher fees.
Initially, the client hoped that using ChatGPT to write the copy would save them a significant amount of money. However, the bland, empty, and commercially uninspiring text not only failed to engage users, but also weakened the brand’s image. He stated bluntly: “At first glance, it’s obvious that it was written by an AI. It’s completely unconvincing.”
These AI-generated texts feature monotonous syntactic structures, rigid rhythms, and a lack of emotion. They are almost impossible to repair and can only be torn down and started over.
Skidd He noticed that the main source of income for many of his colleagues has shifted from content creation to proofreading what AI writes. “Now 90% of the documents provided by clients are written by AI, but we almost always have to proofread them,” admitted one colleague.
Similar AI rework phenomenahave also begun to spread to technical fields, such as development and operations.
In the UK, Sophie Warner, founder of digital marketing firm Create Designs, recently received the highest number of inquiries from clients “duped” by ChatGPT. A customer followed the AI tutorial to modify the code, but the website crashed and was hacked, remaining paralyzed for three days, resulting in a loss of £360.
It wasn’t just small and medium-sized businesses that fell into the trap; large customers were also affected.
Warner said it now charges a “troubleshooting fee” to find AI-related bugs that could have been avoided in advance. As the Chinese proverb goes, drawing a line is worth less than a dollar; Knowing where to track it is worth $10,000.
The original intention of most companies in introducing AI is straightforward: to reduce costs and increase efficiency.
It may seem like a tech dividend not to be missed, but the “State of AI Survey” released in March by global management consulting firm McKinsey & Company shows that last year, 78% of companies used AI in at least one business process, significantly higher than the 55% forecast for 2023. However, the average cost reduction was less than 10% and revenue growth was less than 5%.
Given this disparity, actual usability is secondary; the key is showing others you’re using it. Even if the implementation is rushed and the process is cumbersome, giving the impression of having solved the problem is far more important than actually solving it.
Redazione