ChatGPT still not acceptable for scientific writing. Here’s why
It is a limited study, but it does show the gap that a novice may miss. This will be closed as the companies can retrain and isolate domains…
It is a limited study, but it does show the gap that a novice may miss. This will be closed as the companies can retrain and isolate domains…
https://www.dw.com/en/eu-lawmakers-lay-groundwork-for-historic-ai-regulation/a-65909881 6% of global revenues is a real penalty. GDPR had ripples because of a similar penalty scheme. It is time this approach for penalties is adopted globally…
https://www.theregister.com/2023/07/01/microsoft_github_copilot/ It is interesting to see that they know they are violating licenses by slightly changing code to obfuscate the origin. That kind of deliberate behavior gets full…
https://www.theregister.com/2023/06/28/microsoft_openai_sued_privacy/ First co-pilot causes a suit based on violating open source licenses, therefore copyright infringement. Artists are suing as a class. Now privacy concerns are in the legal…
Bruce Schneier post: https://www.schneier.com/blog/archives/2023/06/ethical-problems-in-computer-security.html
https://www.theregister.com/2023/06/21/openai_government_regulation/
https://www.theregister.com/2023/06/19/even_google_warns_its_own/ Apparently they can’t trust it even though they tell people they can trust it. Wonder what code it is trained on that it can’t be trusted?
Adversarial ML Attack that Secretly Gives a Language Model a Point of View – Schneier on Security
Data ethics for computing education through ballet and biometrics – Raspberry Pi