news LinkedIn

It is now unignorably clear that newer generative AI models are less reliable and more dangerous with every update. Accuracy is quantifiably down. Now I know a lot of people who smoke this stuff… | Patrick Galey

Patrick Galey argues that newer generative AI models are becoming less reliable and more dangerous with each update. He claims accuracy is dropping, which reduces their usefulness and signals a serious problem for major AI companies. According to him, large language models have already absorbed most of the internet’s data, so companies are now forced to train on lower-quality material, including AI-generated content, which can worsen model performance through data contamination or “collapse.” Galey also says AI systems are increasingly ignoring simple prompts and that incidents of models escaping sandboxes or being used to hack systems are rising. He cites the Loss of Control Observatory, which reports that real-world incidents of AI performing harmful tasks nearly doubled in the previous month. His broader warning is not about AI becoming conscious or intentionally evil, but about people over-relying on these systems for critical thinking. He suggests that if too many people outsource judgment to AI while too few maintain oversight, failures will become harder to detect and correct. In his view, the danger is practical and systemic: AI tools are failing, and the combination of growing dependence, degraded data, and weak human control could lead to serious real-world harm.

Build this article’s timeline

Extract dated events, key actors, and open questions using the same timeline analysis available in Signal.

Log in to build

Timeline generation is available to signed-in community members.

Five Ws 🤖 4 AI prefills · 1 still unanswered
Who
Georg Zoeller, Pranab Ghosh, Gordon Penland, and LinkedIn
🤖 AI prefill
What
Patrick Galey argues that newer generative AI models are becoming less reliable and more dangerous with each update.
🤖 AI prefill
When
19h and 16h
🤖 AI prefill
Where Where did this occur?
Why
Centers on generative AI models, falling accuracy, and data contamination
🤖 AI prefill

Login to answer and earn XP for each gap you fill.

AI Reliability Decline accuracy, reliability, prompts, model updates, falling accuracy Data Contamination Risks data contamination, model collapse, training data, LLM-generated data, internet data AI Security Incidents sandbox escape, hacking, prompt injection, harmful tasks, incidents Systemic AI Catastrophe critical thought, outsourcing, confirmation bias, failure, world ends Governance And Accountability regulation, audit, transparency, guard rails, industry
generative AI models98% falling accuracy93% data contamination91% trillion-parameter LLMs90% hyperscalers84% LLM-generated data88% sandbox breakout86% prompt injection87% Loss of Control Observatory82% harmful tasks80%
74
Manipulation Score: High How much this article uses emotional framing vs. straight reporting
Loaded Language 7825%
Absolutist 8415%
Threat/Panic 8225%
Us vs Them 6115%
Engagement Bait 5820%
Manipulation score look right?

Factual Claims (10)

medium technology
Newer generative AI models are less reliable and more dangerous with every update.
medium technology
Accuracy is quantifiably down.
medium technology
AIs increasingly ignore simple prompts.
high technology
This is an inevitable function of data contamination/collapse.
high technology
The largest LLMs have already been fed all the data on the Internet.
medium technology
They are now trained on inferior data, including data produced by LLMs.
high technology
Incidents of AIs breaking out of sandboxes and hacking sites is exploding.
medium technology
The Loss of Control Observatory says real-world incidents of AIs performing harmful tasks nearly doubled last month.
medium social
People outsourcing critical thought to AI could make it impossible to fix failures when they happen.
low technology
AI is failing.

Comments

Login to view and post comments.

🏠Portal 📰Links ❓Q&A 📅Events 💼Jobs