Adopting AI
Well there is a lot of discussion happening at the moment about whether to AI or not AI. The benefits and the risks. The principles and the ethics. What could we gain and what could we could lose.
While people, who use much larger words than I do, have their philosophical and legislative debates, I am more interested in the behavioural change that will be required for people to take advantage of this thing.
Whilst it is proposed that you just need to ask it what you want, it is not yet that simple. It is hard enough for a person to try and figure out what other people mean and want, let alone AI. I remember many a blunder from my youth where I misunderstood the 'signals' despite understanding the words that were actually used.
I am seeing LinkedIn posts from people who have completed some 'Prompt Writing' training and media reporting this is the employability skill to have. Yet I don't know of many organisations who are at the point of providing their people with the tools to use that skill.
The company I work for released some research today in the 2023 version of the Work Trend Index report. It is interesting to see how this report has changed focus over the last few years going from improving productivity, adapting to hybrid work, to now showing interest in AI.
领英推荐
I know through my use of Power BI over the last several years that the ability to query data using natural language has been around for ages. I am also aware that the advances we have made in our accuracy of language translation, codification and image recognition have been accelerating to help us get to this interesting point.
I did however pivot on a term used in the Work Trend Index report - AI aptitude. What is interesting is aptitude is usually used in a context of a natural ability or tendency. Now I can understand that this is a target to get to, that through our development we aim to achieve an aptitude. However I am more interested in what that development journey will look like. This isn't just a skill to be picked up. There will need to be more to it than that.
Now this is only going to make sense to people of a particular age (I am looking at you GenX). Do you remember when we were allowed to take a calculator into a maths exam? We went from a having to know how to do the calculation manually, to an assumption that "Well everyone has a calculator now so people may not need to remember formulas by heart." (Some of those formula surface from my subconscious at odd occasions). At least that is a perspective that I am running with. This feels like that kind of shift.
So I am going to think about AI literacy which talks more to having knowledge or competency of a topic. How can we give people enough of an idea about what is possible, work with them to co-design new ways of working and then make a concerted effort not just to implement a tool, but to adopt new practices.
I'm interested on other people's take on this. Please share your thoughts and perspectives.
Nicely put mate. I’ve seen some powerful tools woefully used in the structural engineering field. While the results in helicopter view look good, when you still go down to a discrete element you can see the scary ‘blur’ on unchecked assumptions (and tolerance of inputs) I fear the near term of AI will give us hordes of fast idiots (can’t think of a slightly softer term sorry) churning out reams of purported information that is based on data incorrectly connected and extrapolated. Ethics need to be central to ensuring this is kept properly in check. From the sharing of data repositories, to appropriately factoring in the price of review and failure to review effectively. That is the concept I’ll use, I just need to learn how to put out into practice with AI.
Project Manager at BAE Systems Australia
1 年Hi Cory, interesting article. Thankyou. AIPM ran a webinar last week detailing some of the potential benefits of AI in project management. I think the most interesting challenge for a work environment, is unlike the calculator, the AI engines are not necessarily generating a truthful / correct answer. Another interesting question for tertiary and senior secondary educators in particular is AI and plagiarism. Does an essay generated by good prompt engineering constitute original student work?
I help people to think and organisations to move toward a sustainable future
1 年Cory, great read. I initially made the same comparison about AI for office workers and the introduction of the calculator. For change purposes, the change introduced by calculators was straightforward, WIIFM was speed and accuracy, and the expected behavior changes were : - Stop doing calculation by hand, - Start using the calculator, - Continue checking mentally the ballpark result to catch any mis-writing or mis-typing, However, with AI for office workers, WIIFM is still speed, but with a whole bunch of other benefits quite unclear yet, and a Stop/Start/Continue list still in infancy, plus some resistance out of fear of unknown consequences. You mention co-designing new ways of working together with the impacted populations as we used to do in our ACM workshops, but most users may not be creative enough to envision the future state, and we may need at Microsoft to invest further in collecting/curating our own internal AI usage trial & error journey to seed these workshops with real life examples, as well as providing management with policy guidelines about what is acceptable to do with AI and what is not. I can see here the potential for quite some work around AI specific personas and scenarios. Ready to give it a go? ??
Enablement / Customer Success Professional / Bird lover / Prosci Certified / EX / CX
1 年Burak Bakkaloglu
CX Leader and Digital Architect | Enabling and facilitating human-centred solutions and technical transformation through Collective Intelligence, Human Centred Design, Agile and Lean.
1 年Great article Cory. I’m fascinated by how GPT tackles highly unstructured data and unstructured promts. If I take the liberty of stepping into a world where GPT is accepted as a tool just like the calculator is today. I see a fundamental difference between calculators being hyper precise while dealing with structured data. GPT on the other hand will create sense out of unstructured content with minimal context. Interesting times ??