Thoughtworks: How to use AI responsibly

It’s been the defining subject of the last few months, and rightly so. You wouldn’t be able to attend a conference in the HR space without having a discussion around artificial intelligence – and if you’re ever wondering which might be the busiest breakout sessions? You guessed it, AI is the topic packing them in.

We’ve all (well, mostly all) acquiesced to the notion that AI is here to stay so the next question is, how can we work with it safely and productively? Mike Mason of Thoughtworks joined editor Jon Kennard to discuss next steps for this groundbreaking technology.

We join the conversation as Mason gives UNLEASH some context about his business.

+++

Mike Mason: I like to tell people we’re the biggest tech firm you’ve never heard of, we’re about 12,000 people. We do a lot of strategy through implementation for our clients. We help them with business strategy, IT strategy, and then actually doing the things: building a lot of software, helping create value for our clients through software.

My role is a technology strategy role at ThoughtWorks. And obviously, with the crazy explosion in generative AI happening over the last few months, that’s been a major focus of my work for the last little while.

Jon Kennard: I was going to say, I don’t know if you’ve heard but it’s been in the news a bit recently, AI, just a little.

MM: Yeah, just a little.

JK: That’s what we’re here to talk about. It’s a landscape, it’s expanding and changing and moving so quickly, in so many different directions. If your organization wants to get involved, or use generative AI in in some way, leverage its incredible capabilities, how do you pivot towards that? How do you get the whole organization on board? How would you convince everyone that it’s the right thing to do, what’s your advice there?

MM: First of all, there’s a significant leadership question here, I think the highest levels of the company need to be making a clear statement about the importance of generative AI, and actually creating organizational structures that support that.

Making sure that the various bits of your organization are tasked with taking this on and figuring out how it affects the business. So a clear statement from the leadership and behavior that embodies that importance. But then the other thing that’s really critical is enabling employees to experiment responsibly on this stuff.

Because most of us have tried some of these tools, whether it’s ChatGPT or one of the image generators, and there’s new stuff every day, and being able to experiment with that stuff with company blessing and guardrails around what is okay to use some of these tools for is really important, because people are only going to learn if they get to experiment.

And it’s actually potentially dangerous for employees to be running around putting potentially sensitive information into these systems.

And so you need to make it clear, what’s a company sponsored, authorized experimentation, responsible experimentation. And the reason that I say experiment is because how you open this, it’s moving incredibly fast.

Something that we learn this month might end up being somewhat out of date next month, and the month after that. So being able to harness the creativity of your organization already, and letting people get to grips with this is really important.

JK: Yeah, definitely. When we were preparing for this and running through ideas, one of the things I thought was really interesting was something you said about the fact that there are two types of jobs that are coming out of this. There’s existing jobs, which are going to be modified because of digital transformation and the move towards adoption of AI. And another whole selection of jobs, which we haven’t even created or even thought of yet, or have to almost scope on the fly about how they’re going to work. So can you give me a few examples of each?

MM: Sure. Existing jobs, basically almost anything that I or you could bucket into the knowledge work category, are potentially going to be impacted by generative AI. That’s actually any decision making that we do. And any creative work that we do. If you imagine someone who’s writing copy for the company’s website.

I’ve done a lot of technical writing with AI helping me and it doesn’t replace me as an author, but it does supercharge what I’m able to do. And I move from being largely creating full complete works myself to fit in an outline of that work, and then having the AI fill that in for me, and then becoming an editor of what the AI is doing.

If you think about other kinds of business, workflow jobs, lots of those are actually ultimately making decisions about things. What is this thing that I’m trying to do within company policy? Can I resolve a customer situation this way or that way? How should I do that? And with AI, we’re going to be able to get some of those recommended decisions coming from AI systems, not always needing to be created from scratch by the humans.

One of the things that you’re going to see is people making many more AI empowered decisions about how they do their job and saying, “Yes, I like that solution that’s been proposed by the AI. No, I don’t like that one because it hasn’t taken into account this other thing that I know is important.”

So, the human value, the value that we bring is all this stuff around context and intuition.

And knowing the situation in our company, and the value we provide to customers and all that kind of stuff. The great thing that AI is now providing is a huge amount of knowledge, these things are reading every webpage on the internet, every book that’s ever been published, and being able to recall that stuff, gives you this superpower of knowledge that you can then apply to your job.

And, in the spirit of us becoming editors and decision makers of AI, you asked about what new jobs might come out of AI. So I asked ChatGPT, and it came up with 15, different potential new jobs in the age of AI. And I liked a handful of them. And I can tell you about them now.

And what I’m doing here is I’m using my own human intuition to filter the stuff that’s coming from the AI. It’s not really replacing me answering the question, but it is giving me some inspiration in directions to go.

One of the suggestions was AI user experience designer. So, when you’re incorporating AI into a system, how does that work itself into the UX? You, as a UX designer, would need to understand how the AI thinks in air quotes, how it responds to certain situations, and then how to represent that in the UI, and then also how to incorporate human feedback into that system over time to make it better.

Another idea was a human/machine teaming manager. So somebody who’s, rather than managing just a set of humans, you’re managing a team of humans and one or more AI systems, and your job becomes to get the most out of that team, which is something that team managers already do. But you’d be adding an AI component into that mix.

And then a third one, there is an AI explainer. So, as AI systems are making more and more decisions, potentially, and giving more and more recommendations, there’s a question about how you understand where and how the AI got to that recommendation or decision.

So, a job role where someone is explicitly understanding the mechanics of how these things are built and trained, and potential biases and problems with them. Actually able to explain that and be a great communicator around that stuff. So AI explainer is another potential role. That’s just a small selection of dozens and dozens that could be generated within seconds as well.

JK: Quite incredible. My final question, the main audience of UNLEASH is HR leaders, we talk a lot about the nexus of AI and HR, how it can change the whole human resources function, make it quicker, more adaptable, better performing? You’d probably say that the CTO and their team is where these things might lie. But how can HR teams lead the way on AI strategy? And how can they support CTOs in their endeavors?

MM: You asked a really great question there. Because this is so broad that it is going to impact all parts of an organization, and so the ability of the HR department to be a good part of that strategic response to this new AI capability is really important.

I don’t think it’s just down to the CTO and the IT organization, although they are probably going to provide foundational AI platform capability and stuff like that, but deploying it and figuring out how does this change our business? How does this let us do something new? Has it let us do something more efficiently? Create more value for customers? All that kind of stuff.

That’s very important that the HR department can work into that. One primary thing that’s really important right now is there’s a lot of fear actually, about how AI is going to make half of us redundant and lead to layoffs and mass change. And I think it’s a valid concern, and one the HR departments should be on top of, and understanding, what are we intending to do here? Can we reassure our employees of the direction that we’re going in, because there is a big question between automating people out of a job versus empowering them to do a better job and to do more.

So I think that’s important, and the change management and cultural change aspects of that are really critical. And then upskilling is the other major area that I call on the HR departments for, because there are going to be jobs, as we’ve touched on, that are created brand new in this.

And there are going to be other jobs where it’s the same job, but you’ve got this AI assistant helping you go faster on it. And there are going to be aspects of that. Things like prompt engineering, how to get the best out of an AI system, how to ask it a question in the right way that can get you the right kind of response to help you go faster in your job and produce a better outcome. That kind of a thing is a new skill. This is going to be rolling out across the organization. So HR teams need to get behind that.

And then the other one, performance management is an interesting question. What does it mean to be doing well at your job now that your you’re AI-enhanced in some way? Is it going to change performance metrics? And again, that comes back to some of the employee comfort questions and making people feel good about the new AI revolution rather than worried about it.

JK: This is a conversation we could potentially have every month for the next however-long, it’s going to change so quickly. But for now, it’s great to get where we are with the latest thinking on this. And in six months’ time, let’s come back, see where we’re at then.

Mike, thanks for your time for speaking to UNLEASHcast.

MM: I’d love to. Thanks, Jon.

+++

You can listen to this interview beneath.