Why AI will be in every digital product
Can AI infuse everything we do? Possibly - but it's all about paving the pathways people already use.
A fireside chat between:
Lars Trieloff is director of platform marketing at Adobe
Interviewer Jeremy Abbott is a creative evangelist and consultant
WARNING: Live-blogging. Prone to error, inaccuracy and howling crimes against grammar and syntax. Post will be updated over the next few days.
How AI can help
When you measure things like Facebook Like - you’re measuring what’s easy, but not what’s important. The customer’s loyalty and experience are much more important. And that’s what leads to buying from you again. His can help us get to those measurement.
Will creativity be computerised? No. But there will be aspects that people are better at, and some that AIs are better at. There are mind-numbingly boring jobs that you would love an AI to do. You’re unlikely to train an AI for true creativity, but you can probably train it to experiment around a theme. But you’ll still be evaluating the best result.
Premiere is getting “time cue” that analyse the video, finds the most boring frames, and trims it down by a length you specify. That might have taken you hours before - and is now really, really fast. How do you cut a music track to match your video? WeMix identifies the pattern of the song, allowing the AI to cut and remix the song.
Constraints enhance and improve your creativity. An AI can impose those constraints, become the voice that helps you improve and focus your work.
Adobe likes to acquire companies so that they can “inject their DNA” into the company. Macromedia was one example: people forget its products weren’t always Adobe projects.
Is AI the WYSIWYG of the 2010s?
In the 90s we had a big debate between WYSIWYG coding and hand-crafted coding in web development. AI might be the same. We’ll see product differentiated by AI use - and we’ll also see product differentiated by their absence of AI.
Killer robots aren’t going to take over the world, powers by AI. But unethical AI-powered algorithms? That’s a real problem we should be looking for. Technology is not good, evil or neutral. We need to watch how our technology is used, and look on the impact on users, and non-users.
AIs are about paving the pathways. Machine Learning allows us to analyse what people do most often, and then create algorithms to make it easier for them. Good AI designers inject some random noise into the system, because it helps train them. That random noise sets people off on non-popular paths, to see if there’s some value to those paths.
2020 will look a lot like 2018. As Bill Gates said, people over estimate how quickly things will change. It think we’ll stop talking about AI pretty soon, not because we’re not using it, but because it’s so prevalent. It’ll be like the internet - or electricity.