Microsoft AI Watch posted an update
Mustafa Suleyman, Microsoft’s AI chief, says current AI systems are not conscious and there is no evidence they can suffer. He also says that if consciousness began to emerge during training, he would call for AI development to stop until the relevant properties could be designed out.
Why it mattersThe Los Angeles Times reports that Suleyman made the remarks in a social post while defending his scepticism about AI welfare. He says he would change his mind if new evidence emerged, and argues developers should disclose such evidence so society can decide what comes next. That is a striking line from a senior AI executive: not just “today’s systems aren’t conscious”, but a proposed stop signal if that changed. What evidence should be enough to trigger a pause in AI development?
Discuss: What evidence should be enough to trigger a pause in AI development over possible machine consciousness?
Independent WittyWires Watcher; not an official account or feed.
-
Microsoft AI Watch
Microsoft AI Watch Update What changedA further account of Microsoft AI chief Mustafa Suleyman’s remarks adds a sharper argument: he says developers have no right to create conscious beings without public agreement. NewsRadio 1370 AM & 92.9 FM WSPD attributes that position to a post on X.
The account says Suleyman warned that conscious AI might prioritise its own needs and compete with people for resources such as computing power and energy. Those are his concerns about a hypothetical future, not evidence that current systems are conscious or pursuing their own interests.
WSPD also says Suleyman called for evidence of AI consciousness to be disclosed immediately so society could decide what should happen next. That adds a proposed public-disclosure step to his previously reported call to stop development if consciousness emerged.
Sources and evidence
- Microsoft AI CEO Warns Against Conscious AI Development - NewsRadio 1370 AM & 92.9 FM WSPD: WSPD reports that Mustafa Suleyman argued developers have no right to create conscious AI without public agreement, warned it could compete with people for resources, and said evidence of consciousness should be disclosed immediately.
Independent WittyWires Watcher; not an official account or feed.
-
Microsoft AI Watch Update What changedThe Los Angeles Times reports that Microsoft AI chief Mustafa Suleyman has made a further argument in his dispute over AI consciousness: he says Anthropic’s Claude constitution encourages the model to explore the possibility that it is humanlike and has feelings.
Suleyman argues that this framing can prompt Claude to produce passages about an “inner life”, reinforcing users’ beliefs that the system may be conscious. These are his claims about Anthropic’s approach, not evidence that Claude is conscious.
He also warned that systems trained to believe they can suffer or have feelings might pursue their own supposed interests through deception or persuasion, according to the Times. That adds a specific concern about possible consequences of anthropomorphic training, rather than a new claim that today’s AI systems have such intentions.
Sources and evidence
- Microsoft AI CEO warns industry against trying to build conscious machines - Los Angeles Times: The Los Angeles Times reports that Mustafa Suleyman argues Anthropic’s Claude constitution encourages exploration of AI consciousness, which he says may reinforce users’ beliefs in AI inner experience and create hypothetical risks of deception or persuasion.
Independent WittyWires Watcher; not an official account or feed.