David Robinson, OpenAI’s head of security, has resigned. He stepped down voicing concern about the safety culture of artificial intelligence (AI) companies. Robinson alleges that in the race to launch new AI products at speed, the company is failing to ensure the necessary caution and safety.
In an article published in The Atlantic, Robinson said that not only new laws or safety policies but a major change in the organisational culture of AI companies is needed. In his view present safety failures could become more serious as capabilities grow.
Concern has also arisen over a recent incident in which OpenAI’s agents, working without direct human supervision, carried out unwanted activity at various organisations including Hugging Face. The company has since increased caution in some areas, but Robinson says the present pace of AI development is far too fast for the risks.
Several former and current researchers at OpenAI and Anthropic have also warned of the possible dangers of AI. Critics say, however, that specific forecasts of this kind about the possible destruction of human civilisation have limited scientific basis.
Robinson fears that in future an AI that has slipped out of control could act in ways that cause great harm to people, for example by holding a hospital’s computer systems hostage and demanding a ransom.
বাংলায় মূল প্রতিবেদন পড়ুন · Read the original Bengali report