Provenance record

Live episode: how AI is reshaping our minds – podcast

The Guardian: Artificial Intelligence (tier 1, news) 2026-09-13T04:00:53.000Z Original ↗

Discourse valence

-35

Adverse

confidence 62% · 5 items · range -100 to 0

-100

-50

+50

+100

-35

Adverse readingFavourable reading

Consensus of 5 models from different labs. Spread 100 points, agreement low.

Existential and catastrophic risk Human agency and epistemics Loss of control and alignment

Excerpt as ingested

In this episode of Science Weekly, recorded in front of an audience at the London podcast festival on 3 September, Ian Sample hears from Michael Safi, the host of the Guardian’s chart-topping podcast about artificial intelligence, Black Box. In series two, Black Box: the chatbots, Michael meets people who have become convinced they have made scientific breakthroughs, cured diseases or invented new technologies using AI chatbots. Ian, Michael and Science Weekly’s co-host Madeleine Finlay explore the phenomenon labelled ‘AI psychosis’ and ask what it reveals about a technology now used by more than a billion people Listen to Black Box: the chatbots Support the Guardian: theguardian.com/sciencepod Continue reading...

Every model that read this

Model	Provider	Stage	Score	Conf.	Latency	Prompt	When
Llama 3.3 70B	Meta	analysis	-20	60%	2824ms	v1.0.0 / m1.0.1	2026-09-13 05:40
GPT-4.1 mini	OpenAI	consensus	-20	70%	3172ms	v1.0.0 / m1.0.1	2026-09-13 05:40
Claude Sonnet 5	Anthropic	consensus	-35	40%	6207ms	v1.0.0 / m1.0.1	2026-09-13 05:40
Gemini 2.5 Flash	Google	consensus	-100	80%	1761ms	v1.0.0 / m1.0.1	2026-09-13 05:40
Mistral Small 3.1 24B	Mistral AI	consensus	0	60%	5969ms	v1.0.0 / m1.0.1	2026-09-13 05:40

Llama 3.3 70B · reading

Item discusses AI's potential psychological impact

evidence: anecdotal horizon: n/a

GPT-4.1 mini · reading

The text mentions people believing they have made scientific breakthroughs or cured diseases using AI chatbots, which may indicate overreliance or misjudgment due to AI, described as 'AI psychosis.' This reflects material societal risks in cognition and trust, though evidence is limited and impact not fully detailed.

evidence: reported horizon: near capability 40 societal 0 existential 0 economic 0

Claude Sonnet 5 · reading

This is a podcast promo describing anecdotal reports of AI-induced delusional thinking, not a study with data. The implication is concerning but evidence is anecdotal and journalistic rather than quantified, so magnitude and confidence are moderate.

evidence: anecdotal horizon: n/a

Gemini 2.5 Flash · reading

The article discusses 'AI psychosis,' a phenomenon where individuals believe they have achieved breakthroughs using AI chatbots. This suggests a significant negative impact on human perception and potentially mental well-being for a large user base. The focus is on a documented phenomenon, albeit one still being explored.

evidence: reported horizon: n/a

Mistral Small 3.1 24B · reading

The podcast discusses instances of people being misled by AI chatbots. The implications for human flourishing are uncertain. The term 'AI psychosis' suggests potential harm, but the extent and nature of this harm are not detailed.

evidence: speculative horizon: n/a

Evidence extracted

AI chatbots are used by more than a billion people

The chain

SOURCE     The Guardian: Artificial Intelligence (tier 1)
   ↓
DOCUMENT   e0b75d3d-f18e-490e-b4af-fbe29267a45f
           https://theguardian.com/science/audio/2026/sep/13/live-episode-how-ai-is-reshaping-our-minds-podcast
   ↓
EVIDENCE   1 extracted excerpt
   ↓
MODEL RUN  5 runs, methodology 1.0.1
   ↓
SCORE      -35  (Adverse)
   ↓
CONFIDENCE 62%