Some design considerations for AI

  1. We need to first develop models where we can co-exist with AI. And, the AIs can co-exist with us. I am not talking about a robot physically sitting next to us. But, we need to be aware of their presence in the form of algorithms, tools or technology. And, they should be aware of our presence by constantly remembering who they serve by aid of interacting via feedback.  
  2. Co-existence with AI is the first step to collaborating with AI. Today, people don't have a good intuitive feel for AI because it feels like magic. It is top down and driven into their lives. It's not a choice. It has to be become a choice to permeate mainstream among the long tail population. It can only become only a choice when people recognize (even at a high level) what the technology can do.
  3. No technology can work like magic. We need to change that perception. Machine learning, AI, NLP, Chatbots, etc, are more buzz words in the industry today in majority of places and people use it as leverage. But we rarely get into an intellectual conversation around what the technology can accomplish and how to design for the technology to be effective.
  4. Current designs and interaction models will fail in a world where AI co-exists among us.
  5. Collaboration with AI is key to developing better algorithms. Today, there is a fair bit of bias. The technology should become self-correcting. As it (AI in the form of product, technology, algorithm, etc) collaborates with more people, it should be able to heal its biases.
  6. Ethics are important. Most of AI algorithms operate on understanding patterns and behavior, which requires a lot of data. The data comes from its interaction with people. We should be open to sharing what data is being used for what. A healthy debate needs to happen among not just the researchers and scientists but the larger community in general,