Architects Find Potential Benefits of AI and Generative Design in Floor Planning
Sorry for the mouthful, but this one's super interesting, while at the same time, complicated (as all the best stories are). We're beginning to see experiments in architectural problems like floor planning through the use of AI, planning software (like CAD) and concepts like Generative Design.
Never heard of Generative Design? Here's Wikipedia's definition:
Generative design is an iterativedesignprocess that involves a program that will generate a certain number of outputs that meet certain constraints, and a designer that will fine tune the feasible region by changing minimal and maximal values of an interval in which a variable of the program meets the set of constraints, in order to reduce or augment the number of outputs to choose from.
The non-wikipedia version:
You decide you want a design made up of circles and squares, you always want them to touch each other. You design something that matches that criteria. This is the basis for computer generated design in it's most simplistic form.
Joel Simon, a computer science grad, design grad and bioinformatics grad (holy triple threat batman) has come up with a very interesting way of utilizing this concept in order to develop highly efficient floorplans that end up looking like something an ant colony would create. Check it out here: http://www.joelsimon.net/evo_floorplans.html
Generative design is not a new concept, in fact it's been a known term for at least 20 years; however, without the emergent access to AI a la IBM Watson or Amazon, it's been a great idea without a lot of juice.
If we take this concept into advertising, we see these ideas showing up in programmatic media buying, where we can customize ad units based on environmental conditions. It's early days, and it's not perfect, but we're seeing platforms that will generate all kinds of permutations of ads that can then be put out into the real world as experiments without requiring giant studios to create every single one.
We've also seen this in Content Management Systems like Grid.io - which autogenerate websites on the fly based on a set number of inputs.
It's early days, but it's important to start thinking about these ideas as they relate to our day to day tasks as marketers, and where we need to place Generative Design, vs. using our human brains.
Fake News Is Evolving Into Fake Events
Earlier this week, Facebook shut down 32 pages that they say were created by Russians with the intent of distributing fake news. The most fascinating outcome of this is that nearly 300,000 people following these pages had signed up for events to rally against issues like white-supremacy. In other words, these fake accounts had amassed huge followings by appealing to the need for activism. These followers were more upset about the pages being shut down, than the fact that they had been duped in the first place.
There's a great write-up of the full story here: https://www.theverge.com/2018/8/3/17646302/fake-events-facebook-russia-influence-campaign-free-speech
Fake Events legitimize Fake News in that they add real people into the mix. Two years ago, fake news was as simple as creating a fake website, making it look like CNN and sharing a link. These hackers are becoming far more sophisticated in their social engineering tactics, enticing real people to add equity to their story through real world event attendance. With real people behind a page, promoting it organically, it's far less likely to think that any of the content is fake.
Eison Triple Thread Develops Stylist AI Using Your Spotify Data
Eison Triple Thread is a fashion brand in San Francisco that follows a similar business model to Frank & Oak. Tell them your size and style preference and they'll send you recommended wardrobe choices every month. The issue has historically been that not everyone knows what their style preference is, resulting in a mismatch between what the brand sends you, and what you're likely to purchase.
Eison is now experimenting with an app that pulls in your Spotify data as a proxy for what your tastes are likely to be. Check out this interview with the owner https://techcrunch.com/video-article/fashionably-ai/
Whether it pans out for them or not, this is a great example of how AI can be used to solve business problems, in a way that has nothing to do with bleeding edge software, or the obsolescence of human beings. It's a modern story of AI in the wild that everyone should pay attention to and learn from - specifically as we seek out new ways of figuring out how to better connect and understand human needs and motivations. AI may have solutions we had never thought about.
Subscribe to Jon Litwack
Get the latest posts delivered right to your inbox