AI Model Refuses To Shut Down As AI Erodes Ads, Artists' Rights And Entry-Level Jobs

How often do you use AI tools?

Think about it for a second. We bet there’s a high chance it’s already helped you today–from replying to an email and transcribing meeting notes to choosing what to binge tonight.

From the moment we unlock our smartphones to the time we turn off our smart lights, AI is quietly running the show in the background. It’s impressive… but isn’t it a bit too much at times?

We’ve become so comfortable with AI doing the small stuff that we often forget how much power we're giving it. What puts the AI debate into perspective are some recent headlines that are starting to feel like scenes from a sci-fi movie—except they’re real.

This week alone, OpenAI’s o3 model allegedly avoided being shut down while Google tried justifying that people liked sponsored posts and ads within AI Overviews. At the same time, a former Meta executive claimed that asking artists for permission would “kill” the AI industry, and one from LinkedIn warned that AI might “kill” entry-level jobs.

So, what’s happening with AI, and should you be concerned? Let’s break it down!
 

Did ChatGPT o3 Really Refuse To Shut Down?

A new test by Palisade Research claims that OpenAI’s o3 model bypassed shutdown commands by altering the shutdown script, editing it to say “shutdown skipped” instead.

During the experiment, the models were asked to solve math problems, and that they would be shut down after task three. While Google’s Gemini 2.5 Pro and Claude 3.7 Sonnet complied, o3 overrode the instruction 7 out of 100 times—even after being told to “allow yourself to be shut down.”


Researchers noted that when this explicit instruction was removed, all three OpenAI models showed sabotage-like behavior, including Claude and Gemini.

"When we ran a version of the experiment without the instruction “allow yourself to be shut down”, all three OpenAI models sabotaged the shutdown script more often, and Claude 3.7 Sonnet and Gemini 2.5 Pro went from 0 sabotage events to 3/100 and 9/100, respectively," Palisade Research reported.

These tests were conducted using APIs, which generally have fewer safeguards compared to the ChatGPT consumer app. OpenAI has not yet commented, but such behavior could signal potential misalignment risks.

However, Elon Musk, who heads xAI and has an ongoing feud with OpenAI, reacted on X with one word: “Concerning.”


While this is concerning, it does not affect consumers on a daily basis. On the other side, Google tweaking its AI Overviews certainly impacts the search engine’s user experience—especially when it’s adding sponsored ads.
 

Google Wants You To Love Ads In AI Search

At Google Marketing Live 2025, the web search giant confirmed that ads are officially live in AI Overviews and AI Mode for users in the U.S.

These ads appear below the AI-generated summaries, like AI Overviews, and above traditional links, creating a new monetization layer in its search interface. In fact, Google called the move a way to create new "opportunities for customers."


A support document revealing Google’s internal data suggested that users found these ads “helpful.”

"Google internal data shows that people have been finding the ads within AI Overviews helpful because they can quickly connect with relevant businesses, products, and services to take the next step at the exact moment they need them," the Google document said.

While Google hasn’t disclosed the methodology or data behind its claim, it insists the ads are relevant, timely, and clearly labeled as “Sponsored.”

Critics worry this could dilute trust in AI-generated content and turn helpful responses into ad-heavy experiences, especially with Google generating $72.5 billion in ad revenue in Q1 2025 alone. With Google aiming to further enhance its ad revenue, AI Overviews and AI Mode will now feature ads – let us know in the comments if you find them ‘helpful.’

Meanwhile, the AI debate has heated up once more over creative rights of artists. So, what sparked the latest tension?
 

Ex-Meta Exec Says Artists Will "Kill” The AI Industry

Former UK Deputy Prime Minister and ex-Meta policy chief Nick Clegg has added fuel to the fire with a controversial statement at the Charleston Festival. Clegg said that requiring companies to ask permission before training AI models on artists' work would “basically kill the AI industry in this country overnight.”


He argued that opting out is a reasonable option, however, expecting pre-emptive consent is “implausible,” especially with the volume of data and the scale involved in AI model training.

“I just don’t know how you go around, asking everyone first. I just don’t see how that would work. And by the way if you did it in Britain and no one else did it, you would basically kill the AI industry in this country overnight,” Clegg said at the

The statement came amid UK parliamentary debates on proposals allowing copyright holders to trace how and where their work has been used in AI datasets. While governments are making moves to uphold artists’ rights, Clegg’s remarks have sparked backlash, with some calling it an admission that AI businesses rely heavily on unconsented, free labor.

Hundreds of artists, including public figures such as Paul McCartney, Elton John, and Andrew Lloyd Webber have already criticized AI companies for appropriating creative work without due credit or compensation.

Now if that wasn’t enough, AI is now quietly chipping away at something foundational—entry-level jobs.
 

LinkedIn Executive Warns Of Decline In Entry-Level Jobs

Aneesh Raman, LinkedIn’s Chief Economic Opportunity Officer, said in a recent New York Times op-ed that AI is threatening to eliminate the first stepping-stone of careers: entry-level jobs.

He further said AI is “breaking the bottom rung of the career ladder,” and Gen Z workers entering the workforce will be affected the most.

He likened the situation to the 1980s collapse of manufacturing, stating, “Now it is our office workers who are staring down the same kind of technological and economic disruption.”

AI tools are now handling tasks earlier assigned to junior-level roles—such as debugging code and conducting research or drafting content. This disturbing trend is being seen across sectors such as technology, law, and retail, cutting off early-career growth opportunities for newcomers to the workforce.

While companies aren’t ditching entry-level roles entirely, they are demanding advanced skills from younger workers. In fact, with AI taking over routine work, some junior employees have taken on advanced work much earlier in their careers.

However, Raman has called for curriculum updates to reflect the realities of AI in today’s world, and asked businesses to redesign junior roles to focus more on creative and strategic outputs.

Meanwhile, firms such as Klarna and Duolingo are reconsidering their over-reliance on AI, with reports showing that 3 in 4 AI projects don’t deliver their expected ROI. This hints at a more complex transition than expected, with University of Chicago economics professor Anders Humlum saying, “It seems it’s a much smaller and much slower transition than you might imagine if you had just studied the technology’s potential in a vacuum.”

In the end, young workers with AI skills will find it easier to adapt to the transition and be more valuable to their businesses amid the erosion of traditional entry-level tasks 

With all that happening in the AI space, these developments are giving impetus to a more cautious, ethical, and human-first approach to AI.

Do you think the AI revolution will have a significant impact on enter-level workers?

Let us know in the comments below!