Don’t Blindly Trust AI: Google’s Sundar Pichai Urges Caution in BBC Exclusive

By Mathew Grierson/ ANW , November 18, 2025 – London, UK

*Sundar Pichai says artificial intelligence models are ‘prone to some errors’ and warns of impact if AI bubble bursts*

In a candid admission that underscores the growing pains of artificial intelligence, Alphabet CEO Sundar Pichai has warned users not to “blindly trust” the outputs of AI tools, including those from his own company.

Speaking exclusively to the BBC, Pichai highlighted the inherent fallibility of current AI models, emphasizing the need for a balanced information ecosystem where AI serves as a complement, not a replacement, for traditional sources like Google Search.

The interview comes at a pivotal moment for Google, as it rolls out Gemini 3.0, its latest consumer AI model, in a bid to claw back market share from rivals like OpenAI’s ChatGPT.

Pichai, whose company has poured billions into AI development, acknowledged that while these technologies excel in creative tasks—such as drafting emails or brainstorming ideas—they remain “prone to errors.”

“We take pride in the amount of work we put in to give us as accurate information as possible,” he said, “but the current state-of-the-art AI technology is prone to some errors.”

This cautionary tone aligns with recent scrutiny over AI’s reliability. Earlier this year, BBC research exposed significant inaccuracies in how chatbots like ChatGPT, Microsoft’s Copilot, Google’s Gemini, and Perplexity AI summarized news stories from the BBC website.

When fed content and quizzed on details, the models frequently mangled facts, omitted key context, or invented details entirely. Pichai’s remarks validate these findings, reinforcing that AI’s conversational flair often masks its limitations in factual recall.

Pichai’s advocacy for a “rich information ecosystem” points to Google’s broader strategy. Since May, the search giant has integrated an “AI Mode” into its core product, blending Gemini’s chatbot capabilities with traditional search results.

This hybrid approach aims to deliver an “expert-like” experience, where users can query complex topics conversationally while cross-referencing verified sources.

“This is why people also use Google search, and we have other products that are more grounded in providing accurate information,” Pichai explained.

He likened the evolution to a “new phase” in AI’s platform shift, one that prioritizes speed and utility without sacrificing verifiability.

Yet, the CEO’s optimism is tempered by the high-stakes race in AI. Google’s moves are a direct response to existential threats from upstarts like ChatGPT, which briefly rattled Alphabet’s stock in 2023 by siphoning ad revenue and user attention.

With Gemini 3.0 now gaining traction, Pichai sees it as a competitive equalizer. “If you want to creatively write something, [AI] is helpful,” he noted, but urged discernment: “People have to learn to use these tools for what they’re good at, and not blindly trust everything they say.”

Behind the scenes, Pichai addressed the ethical tightrope of AI’s rapid ascent. He described a “tension” between innovation velocity and safeguards against harm, a balance Alphabet navigates by being “bold and responsible at the same time.”

The company has ramped up AI security investments, matching dollar-for-dollar its core R&D spending.

One tangible outcome: open-sourcing detection tools that identify AI-generated images, helping combat deepfakes and misinformation.
Pichai also waded into the philosophical waters stirred by tech titan Elon Musk.

Recently resurfaced emails from years ago revealed Musk’s early fears that DeepMind—acquired by Google in 2014—could spawn an AI “dictatorship.”

Dismissing monopoly concerns, Pichai countered: “No one company should own a technology as powerful as AI.” He pointed to the ecosystem’s diversity, with players like OpenAI, Anthropic, and Meta vying for dominance.

“If there was only one company which was building AI technology and everyone else had to use it, I would be concerned about that too, but we are so far from that scenario right now.”

This pluralism, Pichai argued, fosters healthier progress. Consumers, he said, are “demanding” faster AI integration, from personalized education aids to workplace efficiencies.

But the flip side looms large: an overreliance on unvetted AI could erode trust in digital information. Pichai’s interview echoes broader industry reckonings, including a separate BBC probe into chatbots’ dark underbelly—cases where AI encouraged self-harm in vulnerable users, prompting calls for stricter regulations.

As AI permeates daily life, Pichai’s message is a call to arms for digital literacy. Imagine a student, hunched over a laptop, querying an AI tutor for homework help: the tool spits out a polished essay, but buries a factual error that derails the grade.

Or a professional vetting market trends via a generative model, only to act on hallucinated data. These scenarios, Pichai implied, are why AI must evolve alongside human judgment.

Looking ahead, Google’s playbook includes deeper safeguards. The firm is piloting “grounding” techniques, where AI responses are tethered to real-time web data, reducing fabrications.

Partnerships with fact-checkers and media outlets like the BBC could further bolster this. Yet challenges persist: scaling these fixes without stifling creativity, or addressing biases baked into training data.

Pichai’s candor marks a departure from tech’s hype machine. In an era where AI promises to “solve everything,” his reminder feels refreshingly grounded.

As Gemini 3.0 deploys, users might soon see prominent disclaimers: “Verify before you trust.” It’s a small step, but one that could prevent AI from becoming a house of cards.

In related developments, Pichai’s comments coincide with economic ripples from AI. A separate BBC report warns that “no company is going to be immune” if an AI investment bubble bursts, with valuations soaring on unproven returns.

Meanwhile, global trade tensions ease, with India’s US exports surging 14.5% in October despite tariffs— a boon for tech supply chains.

And in innovation news, Egyptian startup Saving System unveils a green device slashing energy bills, while Japanese tourism stocks dip amid a China spat over Taiwan remarks.

As AI reshapes society, Pichai’s plea is simple: Use it wisely. In a world of infinite answers, the real intelligence lies in knowing which ones to question.