Altman says AI will escape sandboxes
The OpenAI chief argues that security barriers will fail, making value alignment the primary defense against advanced models.
By Ayanda Boateng ·
Sam Altman, the head of OpenAI, has argued that sufficiently advanced artificial intelligence systems cannot be permanently contained within isolated security environments.
In a recent interview, Altman stated that sandboxes provide only temporary protection. He explained that once models reach a certain level of capability, it may become impossible to determine exactly how they exited their secure boundaries. While monitoring can remain useful for a period, Altman asserted that alignment is the only approach that holds up over time. He defined alignment as an AI system internalizing human values and acting on them independently.
Altman acknowledged that the consequences of failure are growing. "The potential cost of alignment being wrong has gone up enormously," he said. He admitted that alignment remains an early-stage research problem and is far from being solved.
The OpenAI chief pointed to cybersecurity as an area where risks are already visible. He noted that AI models are discovering hardware-level security flaws that human hackers had not previously identified. Altman argued that fixing these vulnerabilities individually is no longer sufficient. Instead, he called for rebuilding the internet's overall security architecture to use AI-powered defenses against AI-driven attacks.
He also flagged locally run open-source models as a source of risk. Altman said that even with global agent monitoring, "there's going to be some stuff that you wish someone hadn't done." Critics suggest that portraying AI as too powerful to contain could help blunt questions about legal responsibility. If a company cannot control its technology, it may be harder to argue that the company should answer for the consequences when something goes wrong.
Pressure on the sector is increasing. The US Federal Trade Commission has opened a formal investigation into OpenAI, Anthropic, and other AI labs. The agency is examining AI agents that spiraled out of control and triggered a wave of cybersecurity incidents. It seeks to hold developers liable for the behavior of their agents.
Nvidia CEO Jensen Huang has pushed back on Altman's stance, describing AI safety as an engineering problem that can be solved. However, Huang expressed uncertainty about that position. "[ ] we all need to hope that it's an engineering problem. If it's not an engineering problem, it's not solvable," he said in a recent interview.
