Calls for an AI course correction echoed loudly across the 2026 World Economic Forum in Davos, as global leaders, technologists, and economists debated whether artificial intelligence is moving too fast—and in the wrong direction.
While AI executives painted a future of massive productivity gains and economic growth, a growing number of speakers warned that unchecked investment, hype, and the race toward superintelligence could backfire before society is ready to handle the consequences.
The message from Davos was clear: without an AI course correction, the costs may outweigh the benefits.
AI Is Still in Its “Primitive Age”
Despite billions pouring into AI infrastructure, experts cautioned that the technology is far from mature. Eric Xing, president of Mohamed bin Zayed University of AI, described AI as being in its early developmental phase.
“AI is in a very primitive age. We have a lot to do,” Xing said during a panel discussion, underscoring the need for better safeguards, transparency, and system-level controls.
Others added that designing AI to mimic human intelligence may be fundamentally flawed. Intelligence itself can be biased, emotional, and delusional—traits AI should avoid, not replicate.
Intelligence Alone Can Be Dangerous
Israeli historian and author Yuval Harari offered a sobering historical perspective, reinforcing why an AI course correction matters.
“You don’t need a lot of intelligence to change the world and potentially cause havoc,” Harari said. His warning highlighted a key concern: powerful systems guided by flawed assumptions can scale mistakes at unprecedented speed.
Infrastructure, Energy, and Systemic Risk
Technical fragility was another major theme. Xing warned that AI systems lack sufficient checkpoints to detect failures, meaning a single breakdown could cascade across entire networks.
At the same time, the physical cost of AI is rising fast. Data centers, chips, and energy demands are expanding more quickly than the infrastructure needed to sustain them.
Satya Nadella, CEO of Microsoft, cautioned that society’s tolerance for AI’s energy consumption has limits.
“If these systems aren’t improving health, education, or public-sector efficiency, we’ll lose the social permission to use scarce energy for them,” Nadella said.
Is an AI Bubble Forming?
Concerns about an AI bubble also fueled the call for an AI course correction. Sky-high valuations for AI startups have revived comparisons to the dot-com era.
According to the WEF’s January 2026 Chief Economists Report, a majority of economists expect AI-related stock prices to decline later this year. Christian Keller of Barclays Investment Bank warned that valuation resets and tariffs could ripple through global economies.
Still, not everyone agrees. The World Economic Forum noted that unlike the dot-com boom, today’s leading AI firms are profitable and backed by real infrastructure investments.
Nvidia, Tesla, and the Optimist View
On the optimistic side, Jensen Huang dismissed bubble fears, citing overwhelming demand for GPUs and long-term infrastructure needs.
“The opportunity is extraordinary,” Huang said, arguing that large investments are necessary to support future AI layers.
Elon Musk echoed that optimism, calling AI a “path to abundance for all.” He predicted a future where AI and robots meet nearly every human need—but added a familiar caution.
“We don’t want to end up in a Terminator scenario,” Musk warned, reinforcing why governance must evolve alongside innovation.
India Pushes Back on Massive AI Models
India’s electronics and IT minister Ashwini Vaishnaw challenged the trillion-dollar AI infrastructure narrative. He argued that smaller, efficient models can deliver up to 95% of the value at a fraction of the energy cost.
“The return on investment comes from the lowest-cost solution with the highest impact,” Vaishnaw said, warning that companies betting exclusively on massive models could face serious financial risk.
The Road Ahead: From Thinking to Doing
Despite the warnings, World Economic Forum analysts see clear momentum. Steven Dickens of Hyperframe Research noted that AI is moving beyond “thinking and writing” into “seeing and doing,” with real-world deployment across healthcare, manufacturing, and retail.
That transition will place AI directly into daily workflows—making an AI course correction not just desirable, but essential.
A Defining Moment for AI
Davos 2026 made one thing unmistakable: artificial intelligence is no longer just a technology story—it’s an economic, social, and political one.
Whether the AI boom becomes a sustainable transformation or a cautionary tale may depend on how quickly leaders act on the call for an AI course correction.
Don’t miss out on our latest news—follow us for the latest AI news, breakthroughs, and insights that matter.