Spotify's annoying AI DJ and the future of AI in audio
Our last newsletter of the year is a reflection on where AI in audio has come and how crazy it is going to get
I stopped using Spotify to listen to music and podcasts years ago, but after they announced the end of Google Podcasts in September I cancelled my Apple Music and returned… and I was shocked at how the interface was largely unchanged and it was still difficult to find new shows.
The other obvious addition to Spotify is the AI DJ that has a parody Grand Theft Auto voice and does little but explain why it is playing the next set of songs (it is always because you have listened to them or something similar before). Without a feature to be able to talk sensibly back to the DJ (as OpenAI have already rolled out in ChatGPT) Spotify’s “innovation” feels like it has been dredged up from 5 years ago and hitched to the AI rocket going through our culture. Spotify is already using AI to choose songs for you anyway…
My mind has been fizzing the past couple of weeks with where audio is going to go in terms of AI. The Town podcast (which is owned by Spotify) went deep on the business model for audio going forward (on the morning that Spotify fired 17% of their staff). To be fair, The Town handled the issues without favouritism towards their owners, but did admit that big flashy payouts for talent were probably a thing of the past. Audio continues to dominate, but Spotify is gambling on AI rather than Meghan Markle to push them forward next year.
Despite me trashing Spotify’s DJ, I am busy migrating a few pods over to Spotify For Podcasters. It is a great hosting service and, of course, free.
So, as audio, the most lucrative section of new media, reorientates itself and we reflect on how Spotify is not a media company at all, but a tech company it’s interesting to see what Daron Acemoglu, Professor of Economics at MIT, said recently. He said the future holds a best-case scenario where AI “will try to make humans more capable, more knowledgeable, better problem solvers, better able to deal with complex tasks, and as such, will help workers earn a better living, create more good jobs and perhaps even reduce inequality” or there is a worst-case scenario where AI “would damage the economy by laying off a lot of workers because it will go down a very rapid automation path, it will boost inequality between capital and labor and different types of labor”. I think the problem at this point among employees is that there is no trust that most bosses won’t choose the latter.
The focus for AI and audio has been on synthetic voices (including in this newsletter) and there is still an unknown variable around how audiences will respond to total AI vocal domination. However, I do think there is a genuine opportunity to utilise this tech to improve discoverability and distribution. Maybe Spotify should work on that rather than the annoying DJ.
I have toured the continent this year talking about AI and podcasting (Ethiopia, Egypt, Namibia and here in South Africa). It is too early to give details, but I am bullish about a harmonious mixture of people and AI for content creation in 2024. I’ll be detailing how we get on in this newsletter each week and I look forward to hearing your feedback on what we create.
Coding Corner (the gradual process of a journalist learning how to code)
Thanks to The Rundown AI for this tip. I have been experimenting with AI to analyse my X (formerly Twitter) data. You can request your data in the X settings. Then you feed it into the Advanced Data Analysis section in ChatGPT (this feature is only for paying customers). You can then ask it about your life on the platform and how to get better engagement. You can ask it when you should post and even what you should post about to get more followers or likes. I would love to do this every month with the same content to gauge how successful engagement on X is changing in the Musk era.
What AI was used in creating this newsletter?
Only the image above was created using AI (ChatGPT). I promise I will experiment with using AI to create more of this newsletter in the new year.
In the news…
The JournalismAI Festival 2023 happened last week. Congrats to Mallick Mnela (Editor at iHubOnline) and Winston Mwale (Editor-in-Chief at AfricaBrief) for representing Africa by speaking at the conference.
The Pulitzer Center has grants available for journalists to report on how AI is impacting their communities. They want stories that explore with nuance how AI systems are designed, sold, and deployed around the world. Apply here.
Google released Gemini, its answer to ChatGPT. The claim is that it can outmatch GPT-4 on 30 out of 32 standard measures of performance. But, according to the MIT Technology Review, “the margins between them are thin” and by the time it is released GPT-5 will be imminent. Then this story by Bloomberg broke that Google’s video of the tech had been contrived to seem far superior than it actually is.
What’s new at Develop AI?
We are taking a break for the South African summer. We will be back on the 9th of January.
Look out for our new pod (with humans) about the ethics and regulation of AI in the new year. The pod is called “Blocking AI” and should feature a range of guests who are incredibly smart and generally affable.
Also, I would like to extend a huge thank you to everyone who helped Develop AI grow this year: friends, family, my partner, fellow journalists, DW Akademie, Fojo, Bloomberg, The Egypt Media Forum, Podnews, The Podcast Sessions, Jamlab, Media Monitoring Africa and those who were generous enough with their time to guide me through the tech.
See you all next year!
Join our WhatsApp Community, listen to our AI generated podcast, visit our website, or contact us on X, Threads, LinkedIn, Instagram or TikTok.
Physically we are based in Cape Town, South Africa.
You can email me directly on paul@developai.co.za.
If you aren’t subscribed to this newsletter, click here.