Google CEO Sundar Pichai has cautioned users against blindly trusting artificial intelligence tools, emphasizing that AI models are prone to errors and should be used alongside other reliable information sources. His remarks, made in an exclusive BBC interview, highlight ongoing concerns about AI accuracy and the need for a balanced approach to technology adoption.
In the interview, Pichai explained that while AI can be valuable for creative tasks like writing, it is not infallible and often produces inaccurate or misleading information. He stressed the importance of maintaining a rich information ecosystem, where tools like Google Search complement AI by providing more grounded and verified data. This perspective comes as Google integrates AI more deeply into its products, including the recent rollout of AI Overviews in search, which faced criticism for erratic responses.
The company displays disclaimers on its AI tools to alert users to potential mistakes, but this has not shielded it from scrutiny. For instance, Google’s Gemini chatbot and other generative AI products have been known to hallucinate or fabricate answers, raising alarms among experts and the public. Pichai acknowledged these limitations, stating that the current state-of-the-art AI technology is inherently prone to some errors, despite extensive efforts to improve accuracy.
Experts like Gina Neff, a professor of responsible AI at Queen Mary University of London, argue that tech firms should focus on enhancing AI reliability rather than outsourcing fact-checking to users. In comments to BBC Radio 4, she described the situation as companies “marking their own exam paper while burning down the school,” urging greater corporate accountability for AI outputs, especially in sensitive areas like health, science, and news.
Pichai’s warnings coincide with Google’s launch of its latest AI model, Gemini 3.0, which promises advanced performance in understanding diverse inputs like photos, audio, and video. This move is part of Google’s strategy to compete with rivals like OpenAI’s ChatGPT, which has challenged its dominance in online search. The tech giant aims to position Gemini at the core of its products, heralding what Pichai called a “new phase of the AI platform shift.”
Supporting Pichai’s concerns, BBC research from earlier this year found that AI chatbots, including those from Google, OpenAI, and Microsoft, frequently misrepresent news stories with significant inaccuracies. Broader findings indicate that despite improvements, AI assistants still misrepresent news about 45% of the time, underscoring persistent challenges in achieving reliable AI-generated content.
Addressing the rapid pace of AI development, Pichai highlighted the tension between innovation and safety, noting that Alphabet is committed to being “bold and responsible” by increasing investments in AI security. For example, the company is open-sourcing technology to detect AI-generated images, aiming to mitigate potential harms while advancing the technology.
Looking ahead, Pichai emphasized that no single company should control powerful AI technology, expressing confidence in the competitive landscape that prevents monopolistic risks. As AI integration expands, his comments signal a broader industry effort to balance innovation with ethical responsibility, urging users and regulators to collaborate in fostering trustworthy AI systems.
