QUALITY: a price to be paid, or risk a race to the bottom

As technology can create content which has traditionally been the preserve of the human, there is widespread interest in how we now define the quality of content. When machines can match or exceed human technical standards, are the measures and economics of quality still the same?

In this second article of our insight series Human in the Machine, we explore the findings of our surveys, revealing that quality means different things in different contexts, and that in AI-enabled world, technical excellence doesn't mean consumer excellence.

Quality is more than technical excellence; it’s our imperfections too   

Consumers seem to choose content they know has had human input, even when the trade-off is a loss of technical quality associated with that choice.

Our survey indicates that many recognise AI as a new, significant force in the creative industries, with 65% of respondents believing that AI will play more than minor role in how this world develops.

However, while the influence of AI seems inexorable, 56% of respondents harbour reservations about AI-generated content (across all content types) and a significant 49% are uneasy or distrustful of AI-generated non-fiction content. In 2024 the picture was largely the same, at 58% and 53% respectively.

In our article on Truth, we explored how consumers place value on human authenticity, provenance and connection to the human experience. Consumers express concern if this were to be distorted, even as we find that from the outside at least, it can be (increasingly) hard to tell what is human and what is AI-generated.

For those who are yet to be fully comfortable with AI (72%), quality is stated as the decisive factor in what may change their minds, along with the expected benefits of cost and speed. The inference is that AI content is to be held to the same bar as human-created work and so if the output meets the same quality expectations, then consumers could become indifferent or even quick to accept and adopt AI-generated content.

But what is quality in this context?  Our survey delves deeper and explores how people would exercise choices based on explicit trade-offs.  When we look at the answers we see that, for many, quality is far more than a rational or technical measure of excellence. 

These choices are revealing in how people perceive quality and value of content. The technical perfection of something generated via technology isn't necessarily enough to win consumers. People are often willing to accept what might be considered as flaws in exchange for those flaws being borne of a human authenticity and hence defining a uniquely human quality. In short, that imperfection trumps perfection.

However, for many, price trumps principles

There is a striking contradiction emerging. In many ways, respondents tell us they place high value on human qualities in their content.  But when push comes to shove, there is a reluctance to pay extra for that.

While many respondent answers initially point to a prioritisation of human ethics or authenticity, these principles often take a backseat when the conversation turns to cost.

Only 25% of those who prefer human-created content would pay more to protect human jobs (albeit this was only 18% in 2024), and a significant 66% would not pay extra for content if they knew it was created entirely by humans (vs. 68% in 2024). Even among those who value human creativity the highest, 44% would be willing to consume AI-generated content if it was of comparable quality and lower cost.

So although many respondents answers imply that strong principles may be guiding their answers, a reality is that, for many, price ultimately may drive decisions. Consumers may prefer human-created content for its qualities and imperfections, but they may be reluctant to pay a premium for it. 

As technologies that automate manual processes only become efficient at scale, the extent to which consumers are willing to pay for the human role they claim to value becomes a critical factor in the success of AI investment cases.

Looking at other changes between our 2025 and 2024 surveys:

• The demand for ‘human-equivalent quality’ dropped from 34% to 26%
• Those expecting to pay the same for human content dropped from 58% to 49%
• Willingness to pay more for AI-driven personalisation increased from 8% to 10%
• Yet respondents who would "never accept" AI content increased from 22% to 25%
• And factors such as affordability (down 3%) and speed (down 2%) became less important to consumers

What's particularly interesting is how this mixed picture may reflect an immature, or perhaps bifurcating, relationship with AI-generated content. The increase in both those who would ‘never accept’ AI content but also those willing to pay more for AI benefits may be taking us towards a potential polarisation in attitudes between groups. Or perhaps we are now engaged in a more sophisticated and nuanced understanding of where AI has value and benefits?

We’re continuing to analyse the data, and welcome your contributions to the debate.

Conclusion: AI means revisiting Quality vs. Cost as a conversation, both for consumers and companies  

Our survey reveals that quality in AI-generated content exists in a more complex (and perhaps paradoxical) framework than simple technical excellence. While AI can be technically superior, consumers place greater value on those aspects of quality that are human, even when those are manifest as flaws or ‘lower quality’.

However, quality and choice is often reconciled by cost. And while consumers may prefer human-created content, there is a seeming reluctance to pay a premium for it.

Industries and companies are faced with a very real challenge, not just to set and execute strategies for the optimisation of this cost-quality equation, but to make the right - often large - investments in AI and automation technologies, and the associated organisational changes. 

These may, by the very nature of how they need to pay back, force companies towards lowest cost content strategies at the expense of the human attributes consumers may have chosen and valued more highly.  In other words, that AI investment becomes self-fulfilling in creating an economic ‘race to the bottom’ on cost vs quality.

The other way to solve the equation is with consumers themselves.  Can they better understand and differentiate what they value on content so that the choice becomes one between different products?  We don’t expect shoes from a thrift store to be of the same quality as those from a high-end, artisan cobbler. They are both shoes, yes, but the products are unambiguously and unquestionably different. 

The same is true with news gathered by expensive-to-run international network of bureau, trained journalists and technical crews, versus ‘news’ that was scraped online and re-posted or tweeted on social media, transformed by algorithms designed to elicit a scroll from the reader.  As consumers express preferences like those we see in the survey, so educating on where these choices are comparable versus are not, and where products are comparable or not, becomes a key challenge facing content providers.

Addressing questions of quality becomes as critical for strategists, creators and product leaders as it does for CFOs and CTOs:

• How are you investing in technology that preserves quality, as your audience defines it?
• Are you educating your customers on the true value of your services and what that really costs to deliver? If they knew more, maybe they will understand and choose to pay more. 
• How are you balancing your investments in AI to ensure you avoid expensive, hyped or scaled bets that may reduce your strategic flexibility in the future?