Have we finally come full circle with artificial intelligence (AI)? Probably not, at least not yet. But there is something of an awakening, a kind of small market self-correction, that highlights the need for both organizations and nations to rethink how they approach AI.
In the early days, Current iteration of AIThis means an accelerated evolution that began when generative AI (GenAI) and large-scale language models (LLM) came into the spotlight, and is still changing rapidly. Technology vendors assured the world that this technology would not lead to job cuts.
And when layoffs do occur, organizations will strive to explain that the layoffs are the result of resizing, restructuring, or refocusing operations, rather than being caused by AI.
Over time, companies have not been afraid to link AI to layoffs. Still, they took the time to explain why their roles will be impacted by AI implementation, and said some skills cannot be recalibrated, reskilled, or repurposed.
As more time passed, companies became more emboldened and their layoffs affected tens of thousands of jobs at once.
They said AI-driven headcount reduction created greater operational and cost efficiencies than human roles. In fact, this technology islow value of human capital” was no longer necessary…no longer essential…no longer cost-effective.
However, as the number of laid-off and unemployed people increased, unemployment continued, and public perception of AI deteriorated in some quarters. whiplash strike.
Then, companies faced the reality of stepping on the accelerator. Vehicle not fully tested.
Real AI won’t come back because costs will rise
In the last few weeks, Organizations recognize that AI is not perfect And it can lead to costly mistakes such as: Erase the entire production database and Delete the entire codebase.
Forester January Report 55% of employers regret letting go of employees due to AI, and 2 in 3 say they have already rehired, most within six months of the layoff.
The cost savings these companies had hoped for also did not materialize. One-third said they saved money because rehiring costs more than the initial layoffs, but 42% broke even.
That means three quarters got nothing There are financial losses and institutional knowledge losses, Forrester said.
And, to the detriment of their pocketbooks, some have found that it costs much more to spank new AI workers than their human colleagues.
A US CFO reportedly posted $500 million in AI bills after failing to restrict the use of his company’s AI tokens. According to Axios.
Some organizations have since tightened their purse strings.
Uber just exceeded its ceiling. Monthly spending on AI is $1,500 per employee and per agent coding tool, and usage is tracked through an internal dashboard that all employees have access to. Rideshare operators say they can only exceed this limit with permission.
The move comes after Uber revealed in April that it had exhausted its annual AI budget in just four months after encouraging its employees to use AI tools whenever possible.
Amazon too Lowered employee leaderboards This system of tracking staff’s consumption of AI tokens was initially set up apparently to incentivize employees to use AI to perform tasks and move up the ranks on leaderboards. People are now telling us not to use AI just for the sake of using AI.
Amazon reportedly cut 30,000 jobs To direct more investment towards building AI.
Not surprisingly, tight cash has caused businesses to wake up and ultimately rethink their AI deployments.
Simply eliminating the human workforce in favor of AI is not the simple cost-cutting solution that organizations may assume.
However, while AI has proven to be more expensive to ’employ’ than humans, the cost structure may change as the price of computing falls and AI vendors become more ‘innovative’ in how they charge customers for usage.
Jump first, ask questions later
And competition will continue to intensify as companies scramble to overtake the next competitor.
Businesses are already feeling the heat and, in many cases, are being forced to tackle it head-on. without understanding it first What exactly are they jumping into?
A survey of 800 multinational organizations in Asia Pacific, Europe, and the United States found that around 70% are investing in AI, primarily because of its potential or fear of falling behind competitors. Conducted by IDC, the study Requested by Expereo.
One in five, or 20%, admit that fear of being left behind led them to invest in technology with little appreciation.
This number is highest in Asia Pacific at 37%, followed by 13% in Europe and 10% in the US.
Overall, 51% plan to prioritize investments in AI or machine learning over the next 12 months, but only 24% say their AI adoption has exceeded their expectations.
These numbers are higher in Asia Pacific, with 61% prioritizing AI investments and 40% indicating that such deployments are exceeding expectations.
54% of respondents in this region said: higher cost than expected While 49% cited insufficient or substandard training data as a reason for the poor performance of their AI initiatives. A further 46% say AI is simply not working as expected.
Globally, 39% are worried that they won’t know the associated costs and ROI when AI is integrated across their business.
Cognitive impairment is a concern
And it’s not just organizations that are losing sight of what actually matters in the gold rush.
Countries, along with their citizens, are rushing to consume everything with AI, raising concerns about its impact on society.
In particular, it is pointed out that introducing AI too early to children may affect their cognitive development. Mohan Kankanharidirector of NUS AI Instituteis also the Deputy Executive Chairman of AI Singapore.
This is an area that the professor believes should be studied more closely and sooner rather than later.
“We want our future citizens to be fully developed in all aspects, physically and mentally. Accessing AI too soon can have negative consequences,” said Kankanhari, who I spoke to on the sidelines of the event. ATxSummit 2026 Singapore.
He cited the calculator, which at the time of its invention sparked calls for its ban as it could inhibit mental growth.
“As a civilization, we thought: [then] that it can harm us [ability] If you give schoolchildren calculators too early, they will not be able to develop mathematical maturity and ideas. But if you delay too long, you could waste years on pen-and-paper calculations that could have been used more productively, he says.
An agreement was then reached and, subject to each country’s final decision, calculators were allowed to be used in classes from the time students entered secondary school onwards.
“It’s not a complete agreement, but it’s still a reasonable agreement,” Kankanhari said. “We need to establish a similar consensus around the use of AI models.”
“This hasn’t happened yet. I don’t think we’ve had enough serious conversations about it,” he said. “We don’t want to give preschoolers access to AI models and permanently damage their development.”
He called for greater caution here to ensure that such access is introduced in a safe manner and without causing harm in the process.
“This is an example of how and where AI should be introduced and disseminated in society in a responsible manner,” Kankanhari said. “I think someone needs to pay attention to this.”
It’s not just our children who should do it before it’s too late.
Ann MIT research They found that “overreliance” on AI-powered tools can lead to “cognitive atrophy.”
They found that participants who used ChatGPT to write short essays had up to 55% less brain activity compared to participants who didn’t have AI to assist them with similar tasks.
Additionally, the “AI group” was unable to remember what they had written or quote from their essays after submission.
Although the MIT study results were drawn from a small sample base of just 54 participants, they suggest there is real cause for concern.
“While LLM offers immediate convenience, our findings highlight potential cognitive costs,” the MIT researchers wrote. “Over a four-month period, LLM users consistently underperformed on a neurological, linguistic and behavioral level. These results raise concerns about the long-term educational implications of reliance on LLM and highlight the need for deeper investigation into the role of AI in learning.”
Developing the next generation of workers
AI’s initial luster appears to be fading, with a growing backlash, particularly among young people and those in the arts, who have expressed disdain for the suggestion that AI could replace human creativity.
For example, students at the University of Arizona Less-than-enthusiastic reaction In former Google CEO Eric Schmidt’s commencement speech, he mentioned the rise of AI.
Kankanhari explained that the students’ protests were an expression of their fears about the impact of technology. “This is a reality and we can’t just hope it goes away,” he said.
Like many countries that have realized over time that the effects of globalization are not all positive and have resulted in the marginalization of some parts of society, the future will require: Recognize that the same applies to AIhe pointed out.
Job insecurity is on the rise and, like other universities around the world, NUS is deeply concerned about the impact AI will have on the next generation of talent, Kankanhari said.
The school is researching how it can address this issue, including how to update its curriculum and what activities students should be involved in to better prepare them when they enter the workforce.
“Am I just going to be replaced by a model? These are legitimate concerns and we need to address them,” he said. “It is not helpful for industry players to promote the latest models of new products every time they are released.”
he pointed out that AI has great potentialmuch of which currently remains untapped, but not everything in the world requires AI.
The first question we all need to ask is whether we need AI to solve a particular problem, he said, urging everyone not to get caught up in the hype.
Technology should be leveraged and supported wherever possible, but not blindly.
“Understand your business, Where AI really delivers benefits“With careful analysis, don’t just deploy AI across the board, but adapt it where it makes sense,” Kankanhari said.
“and, Society as a whole should benefit It’s not just a few, it’s from there,” he added.
Children should also be encouraged to develop deep skills such as reading comprehension, reflection, reasoning, mathematical skills, and the ability to critically analyze information that is thrown at them.
These skills were important a century ago, they are important now, and they will continue to be important another century from now, he said.
Kankanhari warned that those who do not realize that they are becoming constantly dependent on AI and who do not develop any skills have no future to speak of.
“Replacing reading with bite-sized summaries of everything is dangerous,” he further warned.
Humans are still responsible
More importantly, if you lack expertise and simply consume what the AI model produces, Know when you’ll get an answer That’s wrong.
And humans are ultimately responsible.
The same goes for organizations.
