The rapid rise of generative AI has brought significant challenges, much intrigue, and deep thinking. Many ideas, discussions, and learnings about AI have been shared here on LinkedIn. Thank you, LinkedIn network brains trust and the opportunities at work, for enabling me to learn so much and continue learning about AI. Amongst the abundance of swirling thoughts in my brain about AI, here are five thoughts that stand out to me.

 

1. Safety first

On September 5th, the Australian Government released the Voluntary AI Safety Standard and also introduced the mandatory guardrails for AI in high-risk settings proposals paper. This action is in line with the recent EU AI Act, which seeks to define AI systems by level of risk and the governance of these systems required by the developers or individuals that deploy them. I'm glad to see our Government has adopted this approach. Without safety, we are not living, we are just surviving.

2. AI’s weakness is also AI’s strength

AI is excellent at recognising patterns and generating new patterns. For instance, AI can use patterns to efficiently triage hundreds of documents and infer patterns based on its learnings. When inferences tip over to creative generation, things can go strange. AI can be a very confident and stubborn liar. This would be bad if you require precision and accuracy. However, if you are brainstorming, you’d want things to go a little odd to create novel ideas and this would be good. In each case, when we use AI, we must know how far we can trust AI and when we trust our human knowledge, experiences, and intuition. The same reason it is bad makes it good.

3. AI is (costly) software

AI is not human. Humans made AI, humans created the hardware and the software that is AI, and humans vet the data in the data models. There is a human cost to vetting the data. To make the data safe, paid individuals are exposed to the worst of human content to filter the data and train the model. AI also has an environmental cost. Each AI chatbot conversation uses the energy of about one bottle of water. If you think about all the AI chats and embedded use of AI, this is a mind-boggling amount of energy consumption. AI is costly.

4. Me, myself, and AI

Interactions with AI can seem very human-like. Our interface and experiences with AI are through written and spoken language and vision, as in AI using a camera for visual input. AI needs to be more than human-like; it needs to be human-centred. What is the purpose of AI if not for the social good of all humans, including animals, and the environment? With AI, we are the human in the loop. It is human plus AI. Not just human intelligence or artificial intelligence, rather an additive and shared intelligence.

5. Metacognition and AI

What makes us human? Cognition, creativity, emotions, morals, and culture are some of the many human qualities. Our human cognition and metacognition, our ability to think about our thinking and learning processes, is one of the foundations that make us human. Humans are born with a superpower: their brain. With all its brilliance and flaws, to what extent can AI enhance our human metacognition?

I hope these thoughts have helped you wherever you are on your AI journey. Indeed, AI has made me pause and think deeper. Perhaps AI will make us more human than human.