having time to understand what’s happening, how people want to use these tools, and how society can co-evolve is critical.
SAFETY
-
ChatGPT Bias Challenges and User Psychological Impact
By
–
there will be more challenges like bias (we don’t want ChatGPT to be pro or against any politics by default, but if you want either then it should be for you; working on this now) and people coming away unsettled from talking to a chatbot, even if they know what’s really going on
-
Early AI Tools Release: Balancing Empowerment and Serious Challenges
By
–
we think showing these tools to the world early, while still somewhat broken, is critical if we are going to have sufficient input and repeated efforts to get it right. the level of individual empowerment coming is wonderful, but not without serious challenges.
-
Microsoft addresses AI chatbot image and sustainability concerns
By
–
Los problemas de Microsoft de estos días han sido: A) El chat genera respuestas que dan mala imagen.
B) Es insostenible. Y no sé por qué creías que tenía interés en ocultar alguna de ellas, pero creo que las medidas que han tomado buscan solucionar ambos problemas. [6/…] -
Daily message limits and AI response quality constraints
By
–
"cuando se sabe que lo han capado por las respuestas, no por otro motivo" Explícame cómo el límite de 50 mensajes diarios ayuda a lo de las respuestas, por favor.
-
Sama on AI risks and business opportunities ahead
By
–
Le fondateur de #ChatGPT le génial @sama a déclaré que l’IA allait sans doute détruire notre monde mais qu’en attendant il y a de magnifiques entreprises à créer ! Cela vous inspire quoi ?
-
When Language Models Become Trusted Sources of Hidden Truth
By
–
So what happens when people start to believe language models and their interfaces like ChatGPT reveal some 'hidden truth'?
-
Truth Obfuscation, Censorship, and AI Anthropomorphism Risks
By
–
The obfuscation of truth, censorship and lies — combined with our tendency to anthropomorphise interfaces like #ChatGPT. Things could get a little messy.
-
AI Safety Layers: Vulnerabilities and Bypass Methods
By
–
They can also be censored for 'safety' — but their safety layers can easily be hacked. IE, the case of DAN — the unsanitised version of #ChatGPT.
-
ChatGPT Hallucination and Reliability Issues in AI
By
–
We know that applications like #ChatGPT can lie and 'hallucinate'.