Don't let your organization become a #news story with an(other) #AI failure …. Avoiding common problems when embracing AI – a fireside style conversation fresh from the stage! #generativeai See https://
bit.ly/AIOptimise Such a timely conversation here and in
SAFETY
-
Avoiding Common AI Failures in Your Organization
By
–
-
AI-Generated Responses Reliability Concerns Emerge
By
–
Are we done with old-fashioned internet searches? Not so fast. AI-generated responses are not always reliable.
-
Debating the Effectiveness of AI Image Detection Tools
By
–
C’est pas faute d’avoir prévenu que ça servait à rien de faire croire aux gens qu’on pourrait détecter les fakes images alors qu’il fallait mieux insister sur l’esprit critique des gens en leurs apprenants à recouper des sources plutôt que de donner de fausses solutions.
-
Artificial Species as Invasive: Ecosystem Destabilization Risks
By
–
An invented species would always be an invasive one, wherever you put it. Meaning it risks upsetting any existing ecosystem. Risks are destabilisation of ecosystem, extinction and so on.
-
Jailbreaking Self-Censoring AI Models Demonstrates User Control
By
–
Jailbreaking a self-censoring AI model feels like showing it who’s boss.
-
2024 AI Breakthrough: Capability, Safety, and Positive Impact
By
–
Prediction: 2024 will feel like a breakthrough year in terms of AI capability, safety, and general positivity about its potential impact. In the longer term, it'll look like just one more year on an exponential that can make everyone's lives better than anyone's today.
-
AI Existential Risk and Current Problems Collaboration
By
–
To be clear, I do think AI existential risk is a very important problem. I commend all my colleagues working on it. I also think there are many more current problems that a great many of my colleagues, including those working on existential risk, work on. Let us all work together
-
Red-teaming GPT-4 APIs: Fine-tuning, Functions, Retrieval
By
–
4/ Exploiting Novel GPT-4 APIs – performs red-teaming on three functionalities exposed in the GPT-4 APIs: fine-tuning, function calling, and knowledge retrieval.
-
Critique de la rhétorique extrême sur l’IA et l’influence EA
By
–
This is utter madness: “If stopping malignant AI required war between nuclear-armed nations, that would be a price worth paying.” I’m glad @politico is covering the EA cult and their influence.
-
Hallucinations in AI: From Hidden Problem to Research Focus
By
–
It’s pretty cool how hallucinations emerged as a mainstream research topic due to ChatGPT’s ability to say “Sorry, I can’t answer that.” In the past (i.e., a few years ago), AI would always guess, so we would simply compute: accuracy = num_correct / num_examples. Now we are