Photo Credit: Getty Images
OpenAI CEO Sam Altman is urging caution as ChatGPT gains ever more traction. In the inaugural episode of OpenAI’s official podcast, Altman expressed surprise at how much people trust the AI. “People have a very high degree of trust in ChatGPT, which is interesting, because AI hallucinates. It should be the tech that you don’t trust that much,” he said, stressing that the model still produces misleading or outright false information—despite its sophisticated veneer.
He explained that ChatGPT generates responses by predicting the next word based on training data, not by understanding truth as humans do, which makes it prone to “hallucinations”—invented facts that sound plausible. By warning users not to trust it blindly, Altman aims to strike a balance between utility and vigilance: the tool is powerful, but not infallible.
Alongside addressing hallucinations, Altman tackled privacy and monetisation concerns tied to ChatGPT’s evolving features. Persistent memory and forthcoming ad-supported models improve user experience but raise questions about data security and transparency. Altman emphasized that any monetisation, like ads or transaction fees, mustn’t compromise the integrity of responses—“a trust-destroying moment” if they do.
On the subject of evolving technology, Altman admitted a major shift in his stance on AI hardware. Previously, he believed current computers were fit for the AI era—but now he’s reversed course. In a recent podcast appearance, he noted, “Current computers were designed for a world without AI,” signaling that ordinary screens and keyboards may soon feel outdated. He envisions “third-device” concepts more aware of environment and user context—not necessarily screen-based—currently being explored with other tech collaborators.
Altman also touched on deeper social implications of AI. As a new father, he’s used ChatGPT for baby-care guidance, acknowledging how helpful it can be, even as it skews facts at times. He believes future generations might come to see AI as a natural reference point—their cognitive benchmarks. But with that, he warned, society will have to figure out new guardrails, especially around parasocial relationships and dependency.
Altman’s message is clear: while AI like ChatGPT can profoundly enhance capabilities—from parenting help to content creation—it remains fallible. As we transition toward AI-first interfaces and devices, transparency and critical thinking become more vital than ever.
