AI Backlash: Companies’ Warnings, Scale AI & King Layoffs & WeTransfer’s Woes

As artificial intelligence (AI) technology continues to drive how businesses enhance their operations and expand their services to boost productivity, streamline workflows, and reduce turnaround time, there are numerous concerns that arise with its widespread adoption.

One of the primary causes for concern includes layoffs, which have been rampant since the introduction of OpenAI’s ChatGPT in late 2022.

However, the past few days have seen mass layoffs rear their ugly head again.

Microsoft revealed around 15,000 layoffs due to AI, coming in two parts—one of 6,000 and the other of 9,000, which transpired more recently. At the same time, Xbox’s Executive Producer Matt Turnbull suggested employees replaced by AI to use AI to overcome their unfortunate replacement—in a rather tone-deaf move.

Ahead of this, workers of satellite navigation and location technology company TomTom, Recruit Holdings—the parent company of Indeed and Glassdoor—and others, were also affected.

Now, the next lot of layoffs have been announced.

Renowned data-labeling company Scale AI said it would be dropping 14% of its workforce, or 200 employees, as it looks to streamline operations. The layoffs include 500 of its global contractors.

This move comes almost exactly a month after the company received a huge investment from Meta, which included an exchange of a 49% stake in the company for around $14.8 billion.

“The reasons for these changes are straightforward: we ramped up our GenAI capacity too quickly over the past year,” wrote Jason Droege, CEO of Scale AI, in an email to all Scale employees.

“While that felt like the right decision at the time, it’s clear this approach created inefficiencies and redundancies. We created too many layers, excessive bureaucracy, and unhelpful confusion about the team’s mission. Shifts in market demand also required us to re-examine our plans and refine our approach.”

Droege believes the move will help the company to be more adaptable to market shifts, serve existing customers, and win back customers.

“We remain a well-resourced, well-funded company,” Droege wrote.

Next on the list of recent layoffs is King, the developer of the popular mobile game Candy Crush Saga.

Here, the developer is expected to lay off north of 200 employees—but the move isn’t simply just dropping people.

The laid off workers are said to be replaced by the very AI tools and models they spent years building and training to make their jobs easier. This includes several narrative, UX, level design, and user research employees. Even the copywriting team is in the same boat.

“The fact AI tools are replacing people is absolutely disgusting but it’s all about efficiency and profits even though the company is doing great overall,” said one employee. “If we’re introducing more feedback loops then it’s crazy to remove the developers themselves, we need more hands and less leadership.”


Among all the good news for organizations (at the cost of bad news for their employees), AI is causing them problems as well.

File sharing company WeTransfer was accused of using data from the files they host to train AI models. This came courtesy of changed terms of service, which was announced on July 1, 2025, and are expected to come into effect from August 8, 2025, a move that resulted in some users considering it to be strategies to gain access to training data.

Now, the company has published a clarification stating that it does not use files uploaded for such purposes—a blog post that dives into details.

“We do not claim any ownership rights to the Content. You or your licensors own and retain all right, title, and interest, including all intellectual property rights, in and to the Content,” reads an excerpt from WeTransfer’s blog post.

A spokesperson reiterated the clarification, saying, “We don't use machine learning or any form of AI to process content shared via WeTransfer, nor do we sell content or data to any third parties.”

Furthermore, to reduce confusion, the company has updated the terms in language that’s easier to understand.

Amid all this, researchers from OpenAI, DeepMind, Anthropic, and Meta, along with others, have banded together to sound a siren about AI tech.

Through a published study called “Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety,” the group of over 40 researchers has warned that we might lose the ability to monitor AI reasoning, or how they think, as models evolve.

Currently, the ability to see how models “think out loud” offers some level of transparency, ensuring another layer of AI safety.

However, as models become more powerful, they might stop using human-readable reasoning, opting for internal processes that are faster and more efficient. In fact, researchers have already observed that some models choose to abandon English for unintelligible shortcuts.

As the models evolve, they might not “think out loud” the way they’re doing now, which becomes a bigger problem as they still make questionable decisions, exploit weaknesses, or exhibit misaligned goals.

As a result, the researchers are calling for safeguards, including standardized transparency evaluations and robust monitoring techniques.

“All monitoring and oversight methods have limitations that allow some misbehavior to go unnoticed. Thus, safety measures for future AI agents will likely need to employ multiple monitoring layers that hopefully have uncorrelated failure modes,” concludes the study.

“CoT [Chain of Thought] monitoring presents a valuable addition to safety measures for frontier AI, offering a rare glimpse into how AI agents make decisions. Yet, there is no guarantee that the current degree of visibility will persist. We encourage the research community and frontier AI developers to make best use of CoT monitorability and study how it can be preserved.”

What do you think about the latest developments in the AI world?

Let us know in the comments below!