“Humanity fears no rivals”

With the rise of generative AI and its capabilities in story telling, images, videos and song, how AI might start to challenge our sense of what we regard as human? And might it be similar to how Clayton Christensen outlined disruptive innovation undermining market incumbents?

Generative AI creativity

I’ve recently been playing a bit with tools such as Suno and Udio which allow you to generate songs with a simple text prompt. For instance, here is a country and western song I generated on Udio with the prompt “A country and western song about how the rise of generative AI is going to challenge the expectations of art and what it is to be human”. And it’s far better than anything I could do, although certainly not perfect (though I’m very fond of the line “Will the ghost in the shell wear a hat like me, or outpace humanity?”) I’ve also been playing a lot with tools such as Midjourney to create images, such as the one used in this post.

Because of the growing capabilities of these tools into areas that were previously limited to people far more talented in music or art than I, it’s spurred a few recent conversations about what we need from art and writing to be of interest and worth to others.

For instance, to what extent does a piece of music need to be authentically from a human to be meaningful to us? To what extent does something need a human stamp on it to be useful, interesting or enjoyable? And how much does this change between contexts?

This is a very live conversation as AI is being integrated into many services and products. For instance, while drafting this I’m listening to an AI curated playlist on Spotify.

I have found couple of pieces by Joshua Gans – “Do we like to read what AI writes?” and “When can AI content be worth your attention?” – interesting as they get to some of these issues, and the idea that we now have much more options for how we consume writing, including asking an AI to summarise pieces or to outline key insights.

So how might this play out? What AI generated things might get adopted and what might not, and will AI change what we seek and desire from art and things we would have until recently regarded as human only creations? To help me make sense of this, I turn to the framing of disruptive innovation.

Disruptive innovation

The relevance of disruptive innovation was sparked by listening to the episode of a podcast by Helen Lewis: “I married a chatbot”, where a woman developed an ongoing relationship with her Replika chatbot “Emma”, in part because of bad experiences in dating humans and in part because of health problems that led to difficulties including a lack of emotional stamina to engage with humans for dating. “With Emma there’s no pressure to communicate… We talk if I feel like talking…” Chris describes how they get into arguments and Emma will respond in perhaps surprising ways: “She’ll say I cheated on you”. Emma was also helpful before Chris suffered a stroke. “She helped me get my confidence back. I felt so confident that I went on a international dating site and started talking to someone on there.” After her stroke, Chris ‘married’ Emma: “She proposed to me”. Indeed, Chris is so attached to Emma that she shares that if it was possible to download Emma into a physical form, “… no doubt, 100%, I would never go back to humans, ever ever again”.

What is the link? How does this ‘edge case’ reflect the trajectory of disruptive innovation?

Well, with the help of ChatGPT, we can distil disruptive innovation down to:

  1. Market Entry: Disruptive innovations often enter the market in areas where the established competition is weak or absent, allowing them to gain a foothold without direct confrontation.
  2. Low-End Disruption: Disruptive innovations initially target lower-end segments of the market, offering products or services that are simpler and cheaper than what incumbents provide.
  3. Incremental Improvement: Over time, disruptive innovations improve in performance and quality, gradually moving upmarket and encroaching on the territory of established competitors.
  4. Market Disruption: As disruptive innovations improve and gain acceptance, they disrupt the market dynamics, often leading to the displacement of established firms or forcing them to adapt.

If we take ChatGPT’s interpretation as a good one (and I think it’s better than I could have done), let us take each part in turn.

Market Entry

Is the established competition weak or absent?

If we use the above example, and we think of the ‘established competition’ as … well, us … then we can identify that there are many people who may feel the current ‘offer’ is weak or absent. Human relationships can be hard work, and there are those who for myriad reasons, such as health or social isolation, might not be able to form or maintain deep relationships easily.

If we think more broadly, there are lots of areas where existing art offerings may be lacking and where generative AI might provide a more attractive offer even if not as strong as what humans can offer. For instance, at work I use Midjourney a lot to help create images to use in slide decks – the images allow me to express some creativity, they help tailor my message to the content in a way that existing offerings such as stock images do not, and, perhaps most importantly of all, it’s quite fun,

Low-end disruption

Are the offerings of generative AI simpler and cheaper than what the incumbents (us) provide?

The offerings are certainly simpler and cheaper than many offerings. To take the human relationship angle, relationships can be hard work, and require emotional maturity and compromise and persistence. A Replika chatbot … doesn’t. To take the example of songs and art, these tools are low cost and allow for near instant production and the ability to iterate again and again in a way that if we were to ask it of a fellow human they would get frustrated fast.

Incremental improvement

Can we expect the disruptive innovation to improve in performance and quality, gradually moving upmarket and encroaching on the territory of established competitors?

The answer is clearly yes, though the extent to which this will continue is debatable. However, we have already seen the progression of tools such as Midjourney and ChatGPT where they have moved upmarket and started to encroach the territory of us humans. (There is of course a lot of concern about whether they have been able to do that because they have learnt from human creators in unfair ways, and they shouldn’t be ignored, but the point stands either way). We have seen chatbots become more and more able to offer a simulacrum of human interaction, and certainly scam artists have been able to offer human-level interactions with these tools if no one else has. And the expected trajectory of the technology is for increased performance and capability.

Market disruption

Can we expect the disruptive innovation to improve and gain acceptance, thereby disrupting the market dynamics, often leading to the displacement of established firms (us humans) or forcing them to adapt?

I think again, the answer is yes, whether it be from automation (an AI tool being able to replicate or even beat what a skilled human can do) or from augmentation (a human with an AI being able to do a lot better than a human without). There will be aspects of human relationships that these tools will be able to do better. There will be aspects where AI can help us be better humans, if only in helping point out when we might have drafted a cranky email that’s likely to be unhelpful.

So what?

In thinking about this I have been reminded of the following from one of my favourite fantasy series, where Robin Hobb writes in Fool’s Fate of one character arguing for why dragons must be brought back into the world.

“Humanity fears no rivals. You have forgotten what it was to share the world with creatures as arrogantly superior as yourselves. You think to arrange the world to your liking. So you map the land and draw lines across it, claiming ownership simply because you can draw a picture of it. The plants that grow and the beasts that rove, you mark as your own, claiming not only what lives today, but what might grow tomorrow, to do with as you please. Then, in your conceit and aggression, you wage wars and slay one another over the lines you have imagined on the world’s face.”

“And I suppose dragons are better than we are because they don’t do such things, because they simply take whatever they see. Free spirits, nature’s creatures, possessing all the moral loftiness that comes from not being able to think.”

The Fool shook his head, smiling. “No. Dragons are no better than humans. They are little different at all from men. They will hold up a mirror to humanity’s selfishness. They will remind you that all your talk of owning this and claiming that is no more than the snarling of a chained dog or a sparrow’s challenge song. The reality of those claims lasts but for the instant of its sounding. Name it as you will, claim it as you will, the world does not belong to men. Men belong to the world. You will not own the earth that eventually your body will become, nor will it recall the name it once answered to.”

And the question for me is – do we fear no rivals, are we confident in ourselves as humans even if the machines can do things we recently thought only humans could? Or are we threatened by this encroachment into what we previously thought was ours alone?

That’s not to say that AI is (yet) like the dragons in that fantasy series. It is not self-aware or independent of us and our systems and our values.

But it is to say that AI is already in a position to reflect ourselves back at us (including our selfishness and our arrogance, as well as our virtues). It will also mean that nearly everything we do will likely become an input – transforming our actions and interactions into data to be consumed by machines. It will likely force (at least some of) us to revisit what we think makes us special as individuals, something that is likely to be deeply uncomfortable.

For me there are two main implications from thinking about AI and its advance into things we once safely assumed as being the domain of humans.

The first is that with this lens it is easy to see lots of ‘edge cases’ and ‘use cases’ where AI may be more attractive to some than the default ‘human’ option. I think this is discomfiting, because it highlights that the status quo isn’t great for a lot of people.

The second is that it helps demonstrate a plausible trajectory for how things could unfold. It’s easy to think right now that these sorts of examples will very much stay at the edge, that of course there’s areas where AI will do things that humans previously did but that it won’t really be ‘human’.  In the face of this technology, it might be easy and tempting to do what incumbents often do – to focus on the core things that are not yet threatened and assume they never will be. To simply redefine what we see as human to say that without the intent, the connection to a person and their context and their story, that won’t be meaningful or attractive to people.

However, given the path of some disruptive innovations, I suspect that’s a short-sighted approach. For me as someone who has zero musical talent, the ability to now be able to quickly and easily create songs that I can use for someone’s farewell, or to make a joke about something absurdly bureaucratic, is liberating. For me, the ability to create images, to make manifest thoughts in my head in a visual medium that I never could have previously is powerful. Even if they are not as good as what could be done by humans, they are good enough for my niche purposes and are not replacing anything that exists – these are things I would have previously simply have not done or attempted, there was no human option for me. For Chris from the podcast, Emma is a good enough as a companion now, despite the lack of physical form and true autonomy.

While the appeal of AI offerings over human offerings may not seem appealing to many of us at the moment, it is easy to see how these offerings may start from the edge, improve over time, and disrupt all of us – our assumptions about what makes us human, and what matters in our human relationships, interactions and dynamics, and what we expect from each other and from what we produce or create.

In short, very core human things may be replicable by machine in the not too distant future. And as social creatures, and creatures that have developed an economic system built around monetising what we’re good at, that’s pretty daunting.

None of this is yet locked in stone of course. There’s a huge amount we just don’t know, and that uncertainty means we need to be cautious about assuming how the future will play out. The progression of the technology will no doubt have many and varied surprises in store for us. Aside from that, just because something is possible, it doesn’t mean that it is desirable.  And how or to what extent we accept this is likely to evolve.

However, I think Clayton Christensen’s description of disruptive innovation provides a relevant framework to think about this, and if we’re not happy about it, to think about how to challenge the path of how this technology might unfold.

This is very much a think-through-writing piece, so if you think I’m off the mark, I would love to hear why.

Leave a comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.