Question

When will an agentized-LLM system like Auto-GPT be first reported to have been used to do significant real-world harm?

11 comments11

Total Forecasters43

Community Prediction2024-05-05 (2023-11-30 - 2025-01-13)

Make a Prediction

PDF

CDF

This question is closed for predictions, and is waiting to be resolved

My Prediction

community

lower 25%

median

upper 75%

>2026-12-30

...

2023-11-30 09:08

2024-05-05 20:52

2025-01-13 17:14

4.3%

Authors:

GregorVomScheidt NMorrison

Opened:

May 14, 2023

Closed:

Jun 10, 2023

Scheduled resolution:

Jan 1, 2027

AI Safety

Computing and Math

Artificial Intelligence

DeepSeek advances could heighten safety risk, says ‘godfather’ of AI

Guardian•Jan 29, 2025

DeepSeek’s Hidden Warning For AI Safety

Time•Jan 30, 2025

Generative AI Still Needs to Prove Its Usefulness

Wired•Jan 30, 2025

Learn more about Metaculus NewsMatch

Will an AI system be reported to have successfully blackmailed someone for >$1000 by EOY 2028?

80%

Will an AI system be reported to have independently gained unauthorized access to another computer system before 2033?

95%

Will OpenAI, Google DeepMind, or Anthropic announce that they are pausing all training runs above a certain size for safety reasons, before 2026?

Comments

? comments