Is DALL-E 3 better than Midjourney 5.2? Can any of them generate actual words in images? Can Gemini beat ChatGPT Vision at recognition? Which translator does the best Hebrew to English job? Which AI tool should you bet your business on?
The challenge for media and tech professionals in 2024 and beyond will not be whether to use AI to improve editorial and technical work - it will be how to pick and choose the best one for each use case - at any given point in time. And for companies looking to leverage AI to be competitive, the true edge will come from the ability to navigate the field fluidly and rapidly.
Are you psychic?
Some estimates put the current count at 58,000 AI companies worldwide, a staggering number – the result of an equally staggering growth over a very short period of time, both for incumbents and startups. Nvidia alone grew by 249% in 2023 off the back of AI-powering chips, and the US International Trade Administration estimates the global AI market will grow from $157 billion in 2020 to $300 billion in 2024. And while there are clear structural differences between the dotcom bubble and the AI growth cycle we are in, all previous experience indicates that as the technology matures, 9 out of 10 startups will go out of business. So who do you want to bet your money, your business and your success on? How will you know which ones to gamble on?
Consider this snazzy chart I adapted from Upsilon data ....
...and the answer is of course that you don’t, and you can’t - if you could predict this then you’d be an extremely wealthy owner of Nvidia stock.
Maturing AI services are becoming commodity. On their way there, the field will be constantly shifting with whoever has the highest-quality service - above 95% accurate transcription, most cultural context-savvy translation, most accurate object labelling and so forth - in the lead. And with the speed of AI development, this will not even be a semi-permanent state of affairs - it will continue to change every month, until the field is stabilized with a few, high-quality players for each service. And then these will continue to try to out-innovate one another with other differentiators.
Be like water
So where does this leave the end customer? Every media organization in the world has a 2023 strategy containing the words “Leverage AI” along with some pretty ambitious goals. And it’s long since become clear to everyone that this is a must-have. But the decision to use needs to be followed by a decision to choose. And how will you do that? How will you make a bet on which AI will be the best for the next month, quarter or year and then stick to it?
The answer is of course that you don’t. You don’t make permanent choices - you make temporary decisions, constantly subject to change. This requires a paradigm shift - but it’s not a sudden and expected shift - it is one that we have in fact been on the way towards since the launch of Salesforce two decades ago. In the world of service software, lock-in is counterintuitive.
To apply this logic and at the same time embrace AI services, media organizations have to build workflows that allow for change. This is why we built Mimir to be AI-friendly but also AI agnostic. A modern, cloud-native media asset management system has to be fluid and changeable, because we cannot accurately predict which AI service will be the best in a year from now any more than you can. Media organizations set up their own agreements with AI service providers, all of which integrate with Mimir. The AI services being used can then be changed with a drop down by an administrator - and the only difference the day-to-day user experiences is an improvement in performance.
Choose, don’t lose
Treating AI services as backend components solves another problem - inefficient use of AI in organizations. While experimentation has been the name of the game for 2023, in 2024 and beyond media organisations need to streamline that experimentation into efficient usage. This will never happen if hundreds of journalists have individual accounts in transcription or translation services - they need to be delivered where their content lives, as an add-on service.
As 2024 rolls around everyone has decided to use AI. It’s a fait accompli. Going forward, competitive advantage will be had by choosing the right AI at the right time - and doing so on a flexible basis. It shouldn’t matter which AI service you subscribe to as an organization - you make a deal, or any number of deals, based on quality, or on price, or on whatever differentiator you deem most important. Hopefully you can do so on a consumption basis so switching costs you nothing. Then you make continuous change decisions as they benefit your business - be that yearly, quarterly, monthly or even daily. You get the financial or efficiency benefit, your end user gets the incremental improvement.
It’s a jungle of AI services out there, but with the right machete in hand you can cut to the benefits.
Do you want to learn more about how AI agnostic but still AI friendly asset management and newsroom tools can give you access to the power of AI with less of the hassle, minimizing the risk of vendor lock-in? Please get in touch with us.