The end of news? How Sora fundamentally changes film-making, Open Interpreter blows everyone away, and the national AI race gathers pace

Thought of the week: the end of news?

I was babysitting a group of teenagers last week and did a Vox poll on how many of them read the news. I was unsurprised that none of them read a physical newspaper, however I was a little taken aback by the fact that only one of them read the news on social media. I did a little digging and found some shocking and rather sad statistics about the decline of news journalism. The US, for example, has lost two thirds of its newspaper journalists since 2005. In the UK, local newspapers are almost extinct, with publications like the Birmingham Post, which used to serve 1.2 million people, now selling a massive 844 copies per week. Print media is just plain unprofitable.

OK, so we all know print newspapers are historical artefacts and most people like the convenience of apps like BBC News, but the offline decline in news journalism is only accelerated online. Meta says news now accounts for under 3 per cent of what users see on Facebook. But why does this matter? Well, it matters because we are shifting to a post-truth world of manipulated and monetized attention that demands some guardians of the truth. What’s worse is that, seeing the writing on the wall, almost every major media organisation seems to be trying to licence (i.e. cash out) their content to big tech.

Both the BBC and the Financial Times are now ‘experimenting’ with (i.e. rolling out) AI-assisted journalism. This is being driven in part by a seismic shift in how people want to consume their content. Instagram and social media ushered in a new era where content is pretty, personalised and instant, rather than boring, repetitive and depressing. Why read about death and destruction in Ukraine when you can scroll through endless reels of pretty people doing pretty much nothing?

Sora will fundamentally alter film-making

OpenAI released a new batch of Sora-generated films this week that are mind-blowing in their creativity (I urge you to see them in all their glory above). The mini films are the result of OpenAI’s partnership with a select group of creatives who were given exclusive access to play with Sora. It’s a transparent ploy to keep armies of creatives who will lose their jobs from storming OpenAI’s headquarters and burning down the building. Given that we’re not even at version 0.1, it’s clear that AI will fundamentally alter how films are made in the medium term; in a few years' time, the norm will be entire films generated directly from scripts rather than on-set filming.

The implications are profound, and at once exciting and heartbreaking. On the one hand, it unshackles visual creativity from what has traditionally been a very expensive and time-consuming process. On the other, the traditional filmmaking process, while expensive and laborious, is also what gives visual storytelling its depth, nuance and emotive power; teams of creatives toiling away (often for very little money) to craft unique stories that speak to the human condition.

Sora-like GenAI tools offer cheap, filmic fast food, but they exploit existing copyrighted works as training data without proper attribution, threatening the livelihoods of creatives the world over. The counterargument that AI will create more jobs than it displaces is also debatable – we've seen with voice cloning tech how quickly AI progressed from needing hours of samples to just seconds, reducing the need for voice actors. The endgame is AI that can generate content from scratch with minimal human input.

If media companies think they can cash in by monetising their back catalogue, they should think again. By selling the family silver, they may find the benefits to be short-lived, one-time windfalls that undermine their own businesses in the long run. If AI can create compelling content with self-generated synthetic data, who needs Hollywood? This is happening with the large stock image libraries like Getty. Ultimately, it points to a massive devaluation of all creative content.

The national AI race

It’s official: AI is the new oil. Hardly a week goes by without a government announcing a massive investment in its country’s AI capability. Nowhere is this more evident than in the Middle East. Anticipating a future where their black gold loses its lustre to renewable energy, the rich Arab countries are trying to position themselves as essential players in the AI sector. Last week was Saudi Arabia's turn, with the country announcing a $40 billion AI investment fund, making it AI sector’s single largest investor. Having already bought sport (e.g. the World Cup, virtually every boxing match and Ronaldo), they are using huge tax-free sums and the ability to work without pesky restrictions to entice AI talent to the desert.

This does beg the question of what countries without these resources should be doing and what impact being an information taker rather than maker will have on national competitiveness. Irrespective of resources, every country should have some kind of practical strategy about how to:

• enable – ensure an enabling environment, build infrastructure and find funding

• leverage – dust off those e-government plans and build platforms for AI-driven decision-making

• protect – regulate, legislate and safeguard rights

Open Interpreter unveils the 01 Light

The race to extend AI beyond the little chatbot on your computer is quickening, as more and more companies offer robotic process automating (RPA) software. RPA apps learn how to do stuff on our computers by ‘watching’ what users do and then doing that automatically going forward. This quirky video from Open Interpreter showcases how a physical RPA device can control a laptop via voice commands to do things like read and respond to emails, log on to sites and book concert tickets.

Open Interpreter reminds me of the classic "jacket over the back of the chair" tactic used by employees to give their boss the impression that they're still diligently working at their desk, even when they've stepped away.

Other interesting AI news this week

• See this great article on Politico about the shadowy global battle to tame the world's most dangerous technology.

• Neuralink, Elon Musk’s brain implant company, shared a new video showing its first patient Noland Arbaugh playing Mario Kart using mind control. Musk later tweeted that the long-term goal is to shunt the signals from the brain's motor cortex past the damaged part of the spine to enable people to walk again and use their arms normally.

• The UN General Assembly adopted a resolution to promote safe, secure, and trustworthy artificial intelligence systems.

• More bad new for jobs – an Institute for Public Policy Research report predicts that almost 8 million UK jobs could be endangered due to the adoption of artificial intelligence, particularly affecting women, younger workers and lower earners.

Tools that we're playing with this week

Induced, an AI-powered app that streamlines repetitive browser tasks by teaching it once for the AI Agent to run it smartly every time – with human-like reasoning capabilities.

Alice, a local AI bot that sits on your desktop-based AI platform that interacts with your local apps and executes actions like sending emails, filling in forms, etc.

Suno AI unveiled V3, a text-to-song AI. You can now generate decent two-minute-long pop songs from a single prompt.

That's all for this week. Wishing you all a wonderful Easter break.

Thank you for reading Strategic AI. This post is public so feel free to share it.

So you don't miss a thing, follow our Instagram and X pages for more creative content and insights into our work and what we do.