The journey with Artificial Intelligence (AI) technologies continues to be one that’s undulating, rocky and never dull. From initial wonderment at the new capabilities being available at a consumer product level for the first time (such as personalised audio), to continuing fear, uncertainty and doubt about the ethical implications of some of the systems, there’s rarely a day that passes without AI making the news in some new and interesting way.
It reflects the current maturity of the debate, particularly within the publishing and personalised audio sectors, that things have progressed past the art of the possible and the limitless field of “can we?” with its promise of new horizons and mind-blowing wizardry to “should we?” a question with a much more sensible and design-led grounding that promotes innovation but limits the superfluous, undesirable, dangerous or just plain silly.
The regulation question
There’s a strand of the debate that’s deeply concerned with regulation. We’ve seen it in the political sphere in recent times, not least at major meetings of world leaders. Part of the issue around regulation is about addressing people’s concerns and building confidence, particularly with the wide field of technologies considered under the banner of AI, much of which feels new and unfamiliar.
However, lots of AI technologies aren’t that new. Chatbots draw their origins from the 1960s, generative art AI has been around since the 1970s (or earlier, depending on where you ask), even the incorporation of AI into marketing algorithms for social media companies has a twenty year history.
As such, the technology has been operating within the legal framework governing technology and information governance for quite a while. As with other systems, Data Protection regimes allow people to control and manage their information, with concepts such as the ability to delete, modify and maintain accuracy of information central to that. And increasingly helpful updates have been made to clearly state how AI technologies are impacted and covered.
So when considering the development or adoption of new technologies, the debate around AI and its regulation has served to sharpen the debate about how you best commission and develop technologies that meet user needs, are responsive and accessible, and meet their privacy and trust obligations.
Personalised audio designed for use
In the case of personalised audio, there are legitimate concerns around the process for capturing and reusing people’s voices. There are fears of voice clones being used for nefarious purposes, such as identity theft. It is a clear set of concerns for people in the sector to address.
At Dudley Editions, we have tried to answer those concerns through design choices. A voice belongs to the account of the person it comes from. If you want a book read in your mum’s voice, the route is helping her set up her own account and record for herself, not recording her on her behalf. And what a voice is used for is narrow by design: reading a published book the narrator has chosen, shared with the people they choose.
A trusted voice needs a trusted story
At the same time, there is growing evidence of the therapeutic benefit that the types of connection that personalised audio can broker are a valuable emerging tool for publishers and the whole communications sector to be considering.
The same question applies to what people hear. A familiar voice carries trust, and that trust is only as good as the story it delivers, particularly when the listener is a child. That is why we curate our catalogue carefully and work closely with every author we publish and publisher we partner with, across every stage. Personalisation changes who is reading; the story remains the same.
Asking “should we?” every time
As with all design decisions, we need to be clear about purpose, impact and benefits. We need to show our working and not be afraid of entering and contributing to a complex debate.
For us, that means treating “should we?” as a question we keep asking, not a box ticked at launch. Every new feature, market or partner gets the same test: whose voice is involved, who is listening, and does this bring people closer together or is it simply clever?
We need to get past the lingering vestiges of doing things and developing products just because we can. Better products are purpose-driven, meet user needs, and are grounded in delivering improvements. The power of storytelling to connect communities is fundamental to our ideas of civilisation, throughout history. If we can’t articulate how new technologies can support that ongoing element – building accessibility, widening learning opportunities, delivering inter-generational connectivity – perhaps we should pause and refocus, then consider how better to proceed.
What’s clear is that this debate is drawing continued attention, and that it has a long way to run before reaching consensus, if that is even the desired endpoint. The publishing sector has a strong opportunity to inform and drive the discussion, particularly around the theme of how personalised audio can deliver improved connectivity across our communities.