Some tech leaders have been swift to dismiss predictions that artificial intelligence will spark widespread unemployment, while others believe the impact will be significant. Steve Preston, CEO of Goodwill—a U.S. nonprofit with more than 650 job centers— agrees with the latter, saying the shift is no longer hypothetical, according to Fortune.

He noted that it is already unfolding, with consequences most acutely felt by entry-level and low-wage workers.

Last year alone, over 2 million people sought help from Goodwill’s employment services. Now, Preston says the charity is preparing for a “flux of unemployed young people—as well as other people—from AI.” The 65-year-old, who previously served as the 14th U.S. Secretary of Housing and Urban Development, told Fortune that large organizations are already making “significant layoffs based on a move to AI.”

Call centers and sales teams are feeling the brunt of automation. “I don’t know that it’ll be catastrophic,” Preston explained, “but I do think we’re going to see a significant reduction in a number of jobs. I think it’s going to hit low wage workers especially hard.”

A Blow to Young Workers, Especially Non-Grads

Entry-level jobs, often seen as stepping stones for young workers fresh out of school, are shrinking. Preston notes that the labor market squeeze is particularly painful for Gen Z, especially those without college degrees.

“It’s much harder to find a job,” he said. “It’s really hitting college students right now in the marketplace. It’s really hitting young adults without college degrees.”

Research suggests young men are especially exposed to this wave of unemployment.

This reality contrasts with earlier arguments that skills-based hiring would make degrees less relevant.

“What I’m seeing is of the overall unemployment, people without college degrees have no jobs,” Preston stressed.

And while entry-level jobs once provided the foundations for career growth—mentorship, skill-building, and workplace familiarity—AI is eroding these opportunities. Without that training ground, even higher-level roles could be starved of fresh talent in the future.

U.S. vs. Global Workforce Shifts

The U.S. is not alone in this reckoning. In Europe, unions have warned that AI is hollowing out clerical and customer service jobs, leading governments in countries like France and Germany to expand reskilling subsidies. In Asia, call centers in the Philippines—a major global outsourcing hub—are under pressure, as generative AI begins to replicate tasks previously offshored to human workers.

In contrast, economies like India are attempting to balance the disruption by aggressively promoting “AI plus human” service models, where workers are retrained to supervise or complement automated systems rather than be replaced outright.

The comparative picture highlights the uneven pace of adaptation: while U.S. nonprofits like Goodwill brace for a surge in unemployment claims, other countries are embedding AI literacy and clean tech training directly into national policy.

Preston believes that the path forward lies in skills—both digital and practical. Goodwill has been working with employers to identify what will matter tomorrow.

“Digital skills are really critical,” he said, warning that being active on social media is not the same as mastering workplace technologies.

He pointed to tools like Microsoft Excel, Google Docs, and increasingly AI assistants like ChatGPT and Gemini as gateways to employability.

“People who are proficient in using AI tools are beginning to leapfrog other people going into the marketplace,” he said.

For those outside the corporate track, Preston says clean tech jobs—from solar panel installation to EV charging station maintenance—offer promising alternatives without requiring a college degree.

The warning extends beyond Gen Z. “If you are someone seeking a job in your 30s—or even 40s—and you haven’t acquired those skills, you’re pretty much locked out of a massive percentage of the jobs that are available,” he said.

However, Preston insists it is not too late. He recalls cases of people who went from homelessness to high-paying jobs at firms like Accenture and Google after completing intensive digital boot camps.

“When those people get those skills, we just see the doors busting open,” he added.

In a shift from what it formerly held, OpenAI has acknowledged that hallucinations—instances where artificial intelligence generates plausible but false information—are not only common but mathematically unavoidable, regardless of improvements in training data or engineering techniques.

The admission came in a landmark study published on September 4 by OpenAI researchers Adam Tauman Kalai, Edwin Zhang, and Ofir Nachum, alongside Georgia Tech professor Santosh S. Vempala. The paper, republished by Computer World, establishes a mathematical framework showing why large language models (LLMs) will always produce some level of falsehood, even when trained on perfect data.

“Like students facing hard exam questions, large language models sometimes guess when uncertain, producing plausible yet incorrect statements instead of admitting uncertainty,” the authors wrote.

The finding carries weight given OpenAI’s role in igniting the global AI boom with ChatGPT, which has since been integrated into enterprises, schools, and governments. It also marks a departure from earlier industry claims that hallucinations could be engineered away with better training, fine-tuning, or retrieval-augmented methods.

When Better Models Still Fail

The study demonstrated that hallucinations result from statistical properties inherent to model training. The researchers derived mathematical lower bounds proving that generative models will always retain an irreducible error rate.

Even state-of-the-art systems stumbled on seemingly trivial tests. When asked, “How many Ds are in DEEPSEEK?” DeepSeek-V3, Meta AI’s models, and Anthropic’s Claude 3.7 Sonnet all returned incorrect counts ranging from two to seven. OpenAI confirmed its own systems were no exception.

“ChatGPT also hallucinates,” the paper admitted. “GPT-5 has significantly fewer hallucinations, especially when reasoning, but they still occur. Hallucinations remain a fundamental challenge for all large language models.”

Ironically, OpenAI’s most advanced reasoning models hallucinated more than simpler ones. Its o1 model fabricated details in 16 percent of tests, while newer o3 and o4-mini models produced fabricated results 33 percent and 48 percent of the time, respectively.

“Unlike human intelligence, it lacks the humility to acknowledge uncertainty,” said Neil Shah, VP at Counterpoint Technologies. “When unsure, it doesn’t defer to deeper research or human oversight; instead, it often presents estimates as facts.”

How Rivals Have Positioned the Problem

The study directly challenges the narrative advanced by OpenAI’s competitors. Anthropic, the maker of Claude, has often marketed its constitutional AI framework as a way to reduce hallucinations, emphasizing “alignment” and “trustworthiness.” Google’s DeepMind similarly claimed that retrieval-augmented generation (RAG) could drastically cut hallucinations by grounding answers in external databases. Meta, too, has argued that scaling model size and refining evaluation would push the problem closer to elimination.

But the OpenAI study points in the opposite direction, noting that hallucinations are not byproducts of immature engineering, but consequences of deep mathematical laws. By showing that even rival flagships like Claude and DeepSeek-V3 produced wildly incorrect answers to simple factual questions, OpenAI positioned its research as not just a self-diagnosis but a critique of the broader industry’s optimism.

Flawed Benchmarks, Flawed Incentives

The study also exposed how industry evaluation methods worsen the problem. Current benchmarks such as GPQA and MMLU-Pro penalize models for responding “I don’t know,” effectively rewarding confident but wrong answers.

“We argue that language models hallucinate because the training and evaluation procedures reward guessing over acknowledging uncertainty,” the researchers said.

Analysts say this dynamic is already harming real-world deployments. “Clients increasingly struggle with model quality challenges in production, especially in regulated sectors like finance and healthcare,” noted Forrester’s Charlie Dai.

Permanent Challenge, New Strategies

Experts believe the inevitability of hallucinations calls for a governance shift. “This means stronger human-in-the-loop processes, domain-specific guardrails, and continuous monitoring,” Dai said, adding that existing risk frameworks “underweight epistemic uncertainty.”

Shah drew a comparison to automotive safety regulations. “Just as car components are graded under ASIL standards, AI models should be dynamically graded nationally and internationally based on reliability and risk profile.”

Recommendations include calibrated confidence targets, real-time trust indices for evaluating AI output, and updated benchmarks driven by regulatory pressure and enterprise demand.

A Reality Check for Enterprises

The OpenAI findings echo earlier warnings from academia. A Harvard Kennedy School study found that downstream oversight often fails to catch subtle AI-generated falsehoods due to constraints of cost, scale, and context.

The OpenAI team concluded that the path forward requires industry-wide reform in how systems are tested and trusted, while acknowledging that hallucinations will never fully disappear.

For enterprises, the message is that hallucinations are not an engineering flaw to be patched out, but a mathematical certainty requiring new governance, oversight, and adaptation strategies.

By admitting this publicly, OpenAI not only sets itself apart from rivals who continue to promise engineering solutions but also reframes the debate over AI reliability. Thus, the question is no longer when hallucinations will disappear, but how businesses, regulators, and users adapt to their permanence.