Mustafa Suleyman states on 16 September 2026 that we should not treat AI models as if they possess feelings, preferences, rights, or any entitlement to our welfare. He argues that consciousness underpins our ethical, legal, and political systems, and granting another entity any share of these rights lacks justification. This position appears in his article titled A warning about model welfare, where he suggests such moves will complicate the AI containment and alignment challenge.
The warning targets a growing tendency to anthropomorphise artificial intelligence systems. Assigning moral status to software creates unnecessary friction when developers must prioritise safety over simulated empathy. Suleyman implies that treating algorithms as sentient beings distracts from the practical work of controlling their outputs. The statement reflects a pragmatic view where technical constraints outweigh philosophical speculation about machine sentience.
* Consciousness forms the basis of current ethical frameworks
* Granting rights to models lacks evidential support
* Anthropomorphising AI complicates safety efforts