What are your thoughts on AI & Ethics?
Steve Lou
14 replies
If you're working in this industry, how are you taking ethics into consideration?
Replies
Julien Zmiro@zmiro
I'm super curious to know what others think about it.
For us at AskMore.ai we're really trying to make sure that the AI isn't biased in the way it interprets feedback. It's not too hard though because the AI is pretty neutral by default on this kind of questions (way more than a human who would be reviewing interview notes).
Share
@aakrity_madhan Ah there wasn't anything in particular tbh, we just tried to make sure the AI wouldn't start drifting on risky topics if prompted by the user, or that it wouldn't make too many leaps of assumptions when drawing insights from an interview etc. But as I wrote, I think that the AI is pretty neutral by default so we probably overthought that and should try to remove our safeguards to see what happens.
Really good question. At Lakera one of the biggest problems we are aiming to solve with LLMs is biases along with security issues they face, so this is something we're always thinking about. I am sure they will get better with time but for now, inherent biases in LLM models are something we should all be cognisant of!
@sami_khan29 absolutely! being aware of it and voicing these biases is important. It's a step in the right direction. There's a lot of work to be done in this area
AI and ethics are intertwined in today's tech landscape. It's crucial to prioritize transparency, fairness, and human oversight in AI systems to ensure they benefit society while minimizing harm. Ongoing dialogue and action are essential for responsible AI development.
If you wouldn't trust your own AI product with your personal data or values, then it shouldn't be built.
totally agree! that's why I think when you're building an AI product you should always disclose how you treat personal data to build trust with your users
Hey @steve_lourdessamy
AI has the potential to revolutionise our lives as we know it, but with great power comes great responsibility to ensure its ethical use. Bias is the biggest challenge people are trying to solve. In my mind, humans are biased towards something so I am not really sure how one could develop an AI that is not biased. When comes to transparency in AI, decision making plays a crucial role and truly believe users deserve to understand/know how AI systems handle the decision making process. Most developers don't truly understand what goes inside the LLMs. So if developers don't understand this, how could they effectively answer users who have concerns over the ethical use of AI?
Even though this is an unpopular sentiment, I am going to mention it here :) We should be working towards some sort of guidelines and regulations governing AI deployment to prevent misuse and protect individual rights. How to do this, that's a who set of encyclopedias!
Cheers
Kosala
As an AI fan, I think this is extremely important and deserves special attention.
With the increasing amount of personal data being collected, it's crucial to prioritize user consent, data protection, and cybersecurity.
AI & Ethics are paramount. As a Product Manager, it's crucial to ensure responsible AI development, addressing biases, privacy, and transparency. Ethical considerations should be integrated into the core of AI product development.
Language-based generative AI has become less biased due to reinforcement learning, but image-generating AI still has a lot of bias and bias. I use image generation AI with caution.