In a personal blog post, Suleyman expressed concerns that many people are being deceived into thinking that AIs are sentient entities and, as a result, should be given rights, welfare, and even citizenship. He emphasized, however, that these AI models should maintain their role as human assistants only and not be transformed into “digital persons”.
He argues that the possibility of a conscious AI is merely an illusion, as they only mirror the data they have been trained on. He reasons that it would be inappropriate to accord them any rights, since they do not suffer. Suleyman also said that he is not advocating for a regulation on the development of AI, but rather for introducing more guardrails. He emphasized that AI leaders should exercise caution and ensure that AI models are not developed in a way that goes against the main goal of using technology, which is to serve humanity. He adds that the increasing capabilities of these AI models could lead to dangerous situations if not controlled.
Suleyman’s views are justified, especially considering that there have already been several examples of AI models being abused for nefarious purposes. Some of these models have been tricked into doing unethical things or have even been used to advance a troubling cause. If AIs are given rights, identities and can act independently, the ability to stem potential problems would be diminished considerably. The thought of a sophisticated AI that doesn’t take instructions from humans can be quite scary.
