Key take aways:
- The genie is out of the bottle, and organisations need to adapt and develop skills in AI, and have policies on how it is to be used, especially considering commercially sensitive and confidential information.
- AI and morality is an urgent matter for society to debate and involves a nation-wide conversation.
- Policymakers need to do a lot more to get their heads around it, this includes productivity, data privacy, security, safety, educating the workforce.
- The new generation of workers will be using AI, and businesses/govt will have to adapt by embracing it and ensure it is used well. There was a call from the panelists to ensure school kids are well equipped with AI skills and literacy programs.
- It has a large role to play in improving productivity, by reducing the need for menial tasks, allowing for more time focusing on other tasks.
- Panelists believed regulating the use of AI too harshly such as taxing, may discourage its uptake and slow its development.
- Panelists reiterate several examples throughout history showing the opposite: the internet and the explosion of e-commerce and IT jobs.
- How can AI help achieve better trilemma outcomes? For example: AI can be used to getter better results faster, minimising outage risks by identifying possible points of failure etc.
- AI is a fantastic way to learn a new topic at high-level, as it takes the grind out of learning. Of course, if you are an expert in a certain topic, and ask prompts about such topic, you may easily identify wrong information or at least not a nuanced picture, but for a new mind introduced to a topic, it’s a great way to learn. So, experiment with it, instead of just using google.
- Example on how improve the use and outcome of AI chat boxes such as ChatGPT: Instead of writing a random prompt(s), and picking the first result, you should co-edit. This allows for better results. Continue to add further prompts but treat it like a person “this is good, but add x and y, in the tone of z.” “re-edit paragraph z to include x information without the use of these words.” “While writing x, take y into consideration.” In the event of the AI producing something questionable and dubious, ask it a simple question for how it arrived at such conclusion. It will explain to you how it reached this conclusion, and you can edit out any mistakes or assumptions it made and try again.
- Remaining problem: Sharing of data needs to happen. It is still too hard to access data.