▶The Times They Are A-Changin'

If early stage LLMs are sycophantic, then organisations who have invested in AI are… sycophantic-er?

The slack channel explodes every week; someone has built something for themselves over the weekend. Flames, rockets and claps light up your feed. LLM generated summaries of said 'something' sound good but say little. Then, a senior figure chimes in with a flames emoji and people take notice. They even comment:

'I love this initiative'. More flames emojis follow. Someone even adds a heart, a rocket and a party popper. The excitement amplifies, the frenzy intensifies. You sit and ponder what your contributions to 'AI success stories' are going to be.

But a voice in the back of your head just won't stop pestering you.'What the hell are they talking about?', 'I don't understand'.

To be clear, when people are embarking on side quests and building their new AI powered lives - all power to them. This is a time for the brave and the curious. It's a time for experimentation and ingenuity. If you work in tech and you're not experimenting, you probably should be. I am on that journey myself, just as many other practitioners are too. We all know the lay of the land and we know that to foster an environment of mass adoption (followed by the implicit assumption of a mass boost in productivity) we must be encouraging and optimistic. It would be counterproductive to rain on someone's parade who has likely spent several hours experimenting with AI; leadership is all about empowering others.

The issue for me, is that beneath it all, the same paradigms and problems that we were attempting to solve before AI adoption are still there - at least up to a point. And one of life's original, great questions persists for each and every endeavour:'what is the point?'.

I don't mean this to sound flippant. I am not questioning the point of experimenting with AI but your experiment has to be useful. Not in the sense of roaring success; making 'mistakes' and sharing your learnings is a legitimate cornerstone of science. But the project outcome cannot be considered success just by virtue of the fact that you used AI.

My inner cynic wants to rename these types of Slack channels from 'AI success stories' to 'I used AI'. That little word 'success' may be a positive one, but in this context, it can have a hugely detrimental impact on critical thought. Guess how many challenges or constructive criticisms I have seen in this channel of thousands of posts and 4000 people? I can count on one hand.

Curious about this, I've asked over a dozen colleagues from different disciplines:'what's your view on the 'AI success stories channel'?'. Here are a few responses:

'To be honest, I've muted that channel. Those posts all just feel generic, it feels like spam'

'I want to challenge but I think I'll just look like a Luddite'

'I think a lot of it is nonsense but I wouldn't bother commenting, I don't want to look like a d***'

Despite all the apparent positivity surrounding the channel, the silent majority may harbour significant skepticism; not skepticism of AI, not skepticism of experimentation, but skepticism of every outcome being positive. Certainly, the people that I spoke to feel that constructive criticism and healthy debate are off the table.

I suppose my observation is that whilst an initial challenge to all employees to become AI native was helpful, the total and absolute endorsement of each and every shared use of AI - championed by leadership (intentionally or not) - has created a sycophantic environment where mediocre output is paraded triumphantly and we risk inducing a mass case of Dunning Kruger. There is plenty of evidence of this happening elsewhere also. Many employees are fearful of going against the grain, whilst in some cases, attitudes have descended into apathy. I would argue that jeopardises an empowered environment.

Too many times I've seen bewildering problem statements, reverse written into something someone has built, or a prototype that struggles badly when crosschecked against basic HCI principles. The makers of such items should not be shamed or attacked, we are all learning and we are in this together; but sometimes, a challenge to improve something goes an awfully long way.

I do question whether this is just teething problems; perhaps if you're not in the know when it comes to AI, you may feel that it is not your place to critique others' projects. But the truth is we are all apprentices - relatively speaking. No one has been using AI for long enough to be an expert and therefore I would've thought there must be some acknowledgement that we are likely getting an awful lot wrong along the way.

I think back to product critique sessions from many a year ago and the debate was intense, the ideas were robust and the mental stimulation was palpable (not always, but often). I accept that we are in the infancy of this brave new world, the goalposts have moved and we all must move with them, but I do wonder if we are on the cusp of compromising or even losing something deeply valuable.

My second observation is that people are far more willing to step on others' toes. This is a natural byproduct of the assembly layer disappearing; you can no longer demand others stay in their lane because the lanes all blur into one. If you think about it, this is hardly a new concept; on a granular level, 'UX/UI Designer' was and is a title that has been in the industry for years but for the purists, that's an egregious term - UX and UI are different disciplines within design. Yet perhaps if your title is 'Product Designer', you may simply shrug your shoulders and say 'well I do both'. And in a start up, this is basically the norm: hire a generalist and watch them cook; their day to day will cover all kinds of tasks.

The same thing is happening now, although admittedly in a far more disruptive sense. I've witnessed a push for all product and technology practitioners to adopt an entrepreneurial mindset. If you simply regard that as a call to arms for all of us to take more initiative, then I can only see that as a good thing.

There are risks with this, however, which I would imagine are self-evident. One is that experts will find themselves having new conflicts with people who clearly are not informed enough to debate the topic at hand. Such a dynamic is not new per se but imagine you are a designer of 20 years with exemplary credentials and you have a product manager sending engineers fully coded prototypes without consulting you. You first lay eyes on it only once engineers have started inspecting it and then you spot dozens of UX flaws. Existential questions come into play: who is the designer here? Why was I not aware of this before? What are you doing? Where previously, a rejection of poorly thought out product proposals would be part of your responsibilities, you are now relegated to an afterthought.

With the right governance, maybe this isn't a problem, maybe this literally is the new way of working. An unwavering commitment to old product pyramids and a resistance to change is going to look ridiculous in 2 years. I imagine it will look foolish in 6 months. We are all anticipating a restructuring of product teams (it's already happening); less people will be needed to do the same things. What once required a team of 10, may only require a team of 3. And to be honest, if a product practitioner of an arguably arbitrary title is good at their craft, who's to say they shouldn't step on others' toes? Further still, if we are all on the road to generalism, those toes might not be there to step on much longer anyway.

But 'governance' is a very serious word in a world of such extreme change. If a first risk is alienating experts, then a second is rewarding the wrong people.

Let me tell you a story about two product managers named Rob and Bob. Rob is a doer; they are energetic, proactive and love to take initiative. They are ambitious, motivated and intelligent. They have an inner confidence and they are efficient. They are also uncompromising, lacking in curiosity and have a habit of disregarding others.

Bob is a thinker; they are measured, deeply invested in both data analysis and qualitative research. They are committed to rigorous product methodology and refuse to rush or make wild assumptions. They too, are intelligent but they are willing to defer to others where necessary.

Over the past 12 months, guess who has been rampantly building their own ideas, bypassing designers and engineers. I'll give you a clue: it's not Bob. And which of the two would you trust to leverage AI to make better decisions? I'll give you a clue: it's not Rob.

Therein lies a problem. When you throw down the gauntlet, you may trigger a wave of activity from people like Rob who - while efficient and proactive - are prone to being irresponsible and overestimating their own ability. If Rob is positionally rewarded for their initiative, up to a point you might argue that they deserve it - they've worked hard and have made shipping easier.

But is the product improving? Is the user experience meticulously crafted? Do we really know what the point of Rob's work was? Is Rob-powered AI really better than the team that they are set to replace? It is paramount that leadership don't lose sight of what makes a good product in the first place. Just because people can, doesn't mean they should. Let us not elevate the status of those who are busy just because they are busy. Such behaviour is well documented in the technology world; I can't imagine that it always ends well.

Part of me loves knowing that I can now build a product in an evening. Part of me sits pensively in anticipation of the bad decisions that may be on the horizon.

Conscious that there's a current of criticism with a splash of sarcasm flowing through the previous two subsections, here's a positive observation:

Something that I've always found heartening and enjoyable about working in the design world is that there has never been any shortage of advice to go around. So many kind individuals were willing to directly help me when I was first starting out and there are myriad practitioners around the world sharing their insights and expertise online. It's partly what inspired me to start writing myself.

In a large organisation, however, I've witnessed that things can stagnate somewhat; not specifically among designers necessarily, but among product and technology professionals in general. The downside of working on a singular product in a continuum is that true innovation can suffer and people fall into a trap of incrementalism. Such an environment in turn contains silos, rigid disciplines and minimal change. Not to mention that a large organisation almost always suffers from burdensome layers of bureaucracy. I suppose you could argue that people become too comfortable - learning stutters and the status quo becomes impermeable. Momentum slows and inspiration disappears.

As AI takes root everywhere, such a dynamic cannot hold for long. It shakes the foundations of what people know or thought they knew. When the paradigm genuinely shifts, those with a shred of curiosity within them scramble for clues, answers, resources or tools. And the knock on effect of that has been an improvement in sharing culture. As we sit together on the proverbial tip of the iceberg, many of us seem to embody the implicit understanding that we must share and support one another on our journey into the waters deep.

Despite my quips about an absence of critical thinking in slack channels, my feeling is that we are in a golden age of information sharing. So many colleagues and peers are learning so much. And in the spirit of the design community that I have appreciated for years, all of these learnings are being passed around willingly and enthusiastically - in my experience at least. People are reimagining the entire process and despite the seemingly inescapable fear of the collapse of the economy and our humanoid overlords coming for us, right now I feel like it's worth appreciating all the new knowledge and intelligence at our disposal.