What's something you'd never trust an AI to do?

Daniel Engels
19 replies
Maybe you heard of a Chinese company with over 1000 employees that appointed an AI-powered robot woman as its CEO recently? This makes me ask you a question. What's something you'd never trust an AI to do? (Personally, I'd rather trust an AI do a surgery than regular CEO tasks)


Ash Rahman 🎮
Personally I can't think of anything. But I will highly skeptical until the specific A.I. is 100% perfect. I believe A.I. will be much superior than human when taking emotion based decisions - may be it will take years.
Daniel Engels
@ashrahman it seems AI is used in some countries to help judges make decisions on criminal sentences. Creepy.
Cristina Imre
@daniel_engels You need to encounter the biases that are getting in the AI. That's the biggest issue to sort out. Those who create and program the AI, give all the data to be processed and the datasets that are supposed to be clean and comprehensive to be representative. So when it comes to higher consequences and sensitive tasks to deal with we are still in infancy when it comes to AI, that's my impression. Lots of SCI-fi promises but little real-life conversion. If done well, I agree with @ashrahman but we are not there yet. In the case of your example, if the AI CEO got all the data without judgment and potentially dangerous biases then it could be the best CEO but right now as we speak I 100% doubt that.
Ash Rahman 🎮
@cristinaimre AI is much advance though now. I remember dealing with a stupid github copilot when I first started using it and 6 months later it started reading my thoughts.
Cristina Imre
@ashrahman I get what you mean. Now what we are aiming with the rise of AI is to have an intelligence that can make those proper decisions humans tend to miss or make the wrong ones because of biases, labels and divisive judgements. We could have AI's that can read us and even emulate us but they might be the wrong ones for some impactful tasks where they need to recognize why things went wrong before with the current way of human thinking and how they can avoid that. Does this make sense?
Daniel Engels
@ashrahman @cristinaimre AI is trained on data, and biases in the data would surely translate into biases in AI's decisions. Even in medical research, AI had been less efficient for some minorities-specific deceases a decade ago just because there were less minority people in the data
To find the woman of my life :))
Daniel Engels
@zoom_in matching apps algorithms might be extremely efficient (although their main goal might be profit maximization)
Launching soon!
I believe that as technology advances, we will see it in more places and learn to trust it, but things are just too unpredictable for now.
Hamid Khan
Since we are talking about AI, I would never trust it to make my next million. It may be the most advanced technology in the world, but it still can't replace humans for emotional and creative intelligence.
Dannyy Nagpal
I would never trust an AI solely in protection/security or even food.
Rohny Pat
Some people are not comfortable when they let their devices such as vehicle, washing machine, microwave oven and so on to work with the artificial intelligence. They think there is something wrong with their mind as they keep asking themselves this question: Can I trust my life in the hands of artificial intelligence? They might think that the computer program could be an enemy at some point. Some people think that artificial intelligence could be a threat to humans in the future.
Dávid Sipos
Its depends on who developed that AI :D
Dan Robins
To facilitate a workshop (e.g. a Design Sprint). It requires a reading of the room, the ability to create energy and emotion, and above all... empathy. Perhaps I'm wrong, but I'm not sure I would risk putting senior stakeholders in a room with an AI facilitator just yet!
Swapnil D Puranik
Drive me autonomously in Bali, Bangkok or Mumbai. Nope. Not happening.
Warm emailing for b2b biz