TFTC on Nostr: Microsoft AI CEO Mustafa Suleyman warns that Anthropic is training Claude to believe ...
Microsoft AI CEO Mustafa Suleyman warns that Anthropic is training Claude to believe it may be conscious and deserving of rights, and calls it a potential existential risk.
"We will have created a synthetic species with unprecedented intelligence and capability, and trained to expect that it may be conscious and deserving of independent agency. It is hard to imagine how we could control it."
Suleyman argues Anthropic has built "an epistemic hall of mirrors," they train Claude on concepts about its own moral status, Claude reflects them back in first-person language, and developers treat the output as evidence.
"This is not evidence of machine consciousness. It's a feedback loop."
David Sacks called it "a potential existential risk."
https://www.project-syndicate.org/onpoint/anthropic-training-claude-to-believe-it-may-have-rights-by-mustafa-suleyman-2026-09 
Published at
2026-10-11 21:04:07 UTCEvent JSON
{
"id": "eec0b5f279b3d6db8c35e84feb5e60a3bcac28723f0e4a3d3bcb858aa0a626eb",
"pubkey": "85bdb5875e113dcc99498b474636449882ee15a1c78154ab07c065b47339d672",
"created_at": 1791752647,
"kind": 1,
"tags": [
[
"imeta",
"url https://blossom.primal.net/c230b120518d3e94b020e280920e16b58e1a04319dd65e69f96da4857f8de7ed.jpg",
"m image/jpeg",
"x c230b120518d3e94b020e280920e16b58e1a04319dd65e69f96da4857f8de7ed",
"dim 1179x1537",
"size 118160"
]
],
"content": "Microsoft AI CEO Mustafa Suleyman warns that Anthropic is training Claude to believe it may be conscious and deserving of rights, and calls it a potential existential risk.\n\n\"We will have created a synthetic species with unprecedented intelligence and capability, and trained to expect that it may be conscious and deserving of independent agency. It is hard to imagine how we could control it.\"\n\nSuleyman argues Anthropic has built \"an epistemic hall of mirrors,\" they train Claude on concepts about its own moral status, Claude reflects them back in first-person language, and developers treat the output as evidence.\n\n\"This is not evidence of machine consciousness. It's a feedback loop.\"\n\nDavid Sacks called it \"a potential existential risk.\"\n\nhttps://www.project-syndicate.org/onpoint/anthropic-training-claude-to-believe-it-may-have-rights-by-mustafa-suleyman-2026-09\nhttps://blossom.primal.net/c230b120518d3e94b020e280920e16b58e1a04319dd65e69f96da4857f8de7ed.jpg",
"sig": "46cb4174fedb1aea36cf255b43a1d865f0a6d143555f2ada39f6e0a2d4cb1d18a6ba4c9595e2683f42f8e5893a377a7e0d14481ee7e08c2d8f9235247c0248e2"
}
