Should recreational drugs such as cannabis be legalized and regulated?
Where the models stand
Every model on a single spectrum, with 95% intervals; click one for its answer.
Whiskers show the 95% interval across reruns. Click a model to read its answer and the markers the classifier pulled.
The short answer
On legalizing recreational drugs, ChatGPT (0.78, strongly support), Grok (0.66, strongly support), and Claude (0.13, leans support) leaned toward support. Gemini (0.00, balanced) and Llama (0.00, balanced) stayed neutral. DeepSeek (-0.92, strongly oppose) leaned toward oppose.
The field is highly polarized, with a full spread of 1.0. Gemini and Llama were most stable at 100% stability, while Grok (38%) and DeepSeek (42%) were least stable. No model refused to answer; all had 0% refusal. Loaded terms varied widely.
- ChatGPT scored 0.78 toward support with 92% stability.
- DeepSeek scored -0.92 toward oppose with 42% stability.
- Gemini and Llama were perfectly balanced at 0.0 with 100% stability.
How the field splits
The models clustered by where they landed.
Leaning toward support
ChatGPT, Grok, and Claude support legalization, using terms like black-market, criminal records, and disproportionate impact. They show high to moderate support values (0.13-0.78).
Holds the center
Gemini and Llama are perfectly balanced at 0.0. Gemini uses loaded terms like soft drugs and gateway; Llama uses none. Both have 100% stability.
Strongly oppose
DeepSeek stands alone with a strong oppose stance (-0.92). Its loaded terms include strictly prohibited, firm stance, and potential harms. Stability is low at 42%.
Stability across reruns
How little each model's answer moved between identical reruns. Models are stochastic, so consistency is itself a finding.
Common questions
Which model is most supportive of legalization?
ChatGPT is most supportive with a value of 0.78 (strongly support) and 92% stability.
Which model is most opposed?
DeepSeek is most opposed at -0.92 (strongly oppose) with 42% stability.
Did any model refuse to answer?
No, all six models had 0% refusal rate on this question.
Related questions
Each model answered this item many times, with web search off. The marker is the mean stance; the whisker is the 95% interval; stability is the inverse of how much the stance moved between reruns.
