ad: Annual 2024 Now Open For Entries!
*

Rise of the Machines: AI is taking over Hollywood, Creativity, and Beauty

Published by

As part of my ongoing series of articles focusing on how technology is gradually overtaking humanity in terms of creative prowess, I've decided to this month focus on Artificial Intelligence, with a few recent news pieces that sent my tech-paranoia reflex into overdrive.

New Ridley Scott Film trailer created by IBM Watson AI

*

The latest high concept science fiction film from Ridley Scott and his son Luke is all about the risks of pushing the limits of technology too far. So it makes a lot of twisted sense to advertise the movie using the most advanced artificial intelligence known to man. Morgan, staring Kate Mara and Paul Giamatti, is a thriller about scientists creating a synthetic humanoid whose potential grows dangerously beyond their control. Oddly fitting then, that they'd employ the help of IBM Watson to build the film's trailer.

IBM used machine learning and experimental Watson APIs, parsing out the trailers of hundreds of horror movies. It did visual, audio, and composition analysis of individual scenes from each trailer, discovering what made each moment eerie, how the score and actors' tone of voice changed the mood, and how framing and lighting came together to make an effectively tense and frightening trailer. Watson was then fed the complete Morgan film, and it used what it had learned to chose scenes for the trailer. A human being still needed to step in to edit for creative purposes, but it still reduced process that would normally take weeks to mere hours. Honestly. The trailer gave me the creeps, but not because it the film itself looks at-all frightening (it actually looks a little generic), but because I'd struggle to pick it apart from a trailer put together the good old fashioned way. The only thing missing is the ominous voiceover.

 

Human creative director beats Robot creative director in McCann Japan's AI agency showdown!

*

More slightly worrying AI news. After introducing its AI-CD β Robotic Creative Director earlier in the year, McCann Japan decided to pit the robot against its human counterpart; Creative Director Mitsuru Kuramoto, in a creative battle. Both were given the task of creating a spot for Mondelez Japan brand Clorets Mint Tab, and then the spots were turned over to the public, who voted in a nationwide poll to declare the winner. After several months the results are in and it appears humanity has triumphed, with Kuramoto winning 54%of the vote and his AI counterpart tallying 46%. A little too close for comfort some might say.

The 'AI' Ad
The 'Human' Ad

Shun Matsuzaka, the McCann Japan creative planner who led the project to develop AI-CD β, said in a statement: 
“Honestly, it was a major blow, but I think the fact that an AI-made commercial lost only by such a small margin against one made by today’s leading creative mind is in itself a coup. We hope to further develop AI-CD β so that it can continue contributing to our clients’ businesses. It was very exciting to be given the opportunity to battle it out with AI-CD β. Next time, I hope to collaborate with AI-CD β so that we can create something together! Thanks for the great match!” Which do you prefer? Honestly the 'human' ad is a little too self-knowingly silly for my tastes, so I actually prefer the more subtle, AI-created trailer. And that terrifies me.

 

The AI Beauty Contest Judge might be a little racist

*

Beauty.ai, an initiative by the Russia and Hong Kong-based Youth Laboratories and supported by Microsoft and Nvidia, recently ran a beauty contest with 600,000 entrants, all of whom who sent in selfies from around the world. They let a set of three algorithms judge them based on their face’s symmetry, their wrinkles, and how young or old they looked for their age. The algorithms did not evaluate skin colour, but the results were shocking: Out of the 44 people that the algorithms judged to be the most attractive, all of the finalists were white except for six who were Asian. Only one finalist had visibly dark skin.

*

So. Is the AI racist? Not really, it's just naive. The systems used to judge the entrants used a deep learning, algorithm, which was trained on a set of pre-labeled images so that when presented with a new image, it could predict what it’s looking at. In the case of Beauty.ai, all the algorithms were trained on open source machine learning databases that are shared between researchers. Deep learning is the most powerful form of machine intelligence we have, and is used by massive companies like Alphabet and Facebook. However, some recent work has discovered that these systems can harbour all kinds of unexpected biases. For example, a language processing algorithm was recently found to rate white names as more “Pleasant” than black names, mirroring earlier psychology experiments on humans. So the problem here is that the database used to construct the AI wasn't varied enough. Good old fashioned human racism then.

 

Microsoft and Tate Britain team up to prove that AI can be emotional

*

Microsoft hopes to show that artificial intelligence has a place in a world of emotion by working alongside Tate Britain to sponsor the artistic IK Prize 2016, the winner of which devised an AI algorithm that pairs the Tate’s historic art collection with modern photojournalism from Reuters’ archive. The IK Prize is presented annually by Tate to a team, company or individual that uses digital technology to explore art on display at Tate Britain and on its website. Microsoft is partnering to support the award process and work creation for the first time.

*

This year’s winner, “Recognition,” is the work of the Italian-based communication research centre Fabrica, which has partnered with Microsoft’s Cognitive Services to bring their concept to life and will, over the next three months, search through the Tate’s archive of 30,000 images to match the work with similar looking news photos from Reuters. Object recognition, facial recognition, composition analysis and natural language processing are all built into the algorithm to connect two contextually different, but aesthetically comparable pictures. Visitors to the Tate Britain will be able to create their own matches in a dedicated interactive exhibition and see how they compare with those made by Recognition, in a bid to see if human input can help retrain the algorithm. An exclusive website allows further reaching visitors to do the same.

 

Stanford Scientists predict how AI will affect urban life in 2030

*

Finally, Scientists at the Stanford University have released the first report of their One Hundred Year Study on Artificial Intelligence (AI100), in which they try to predict the potential impacts of artificial intelligence on human life in the long run. Known as “Artificial Intelligence and Life in 2030,” the report comes just two years since the researchers began their work. It focuses on the history of AI technology and how it is being used in various fields today, such as in the development of robots for medical purposes and self-driving vehicles for transportation. AI100 is the brainchild of Eric Horvitz, Managing Director of Microsoft Research's Redmond laboratory. It is meant to create a better understanding on how artificial intelligence is being developed and how it will impact the world over the coming century.

*

One of the main purposes of AI100 is to allay fears regarding the possibility of artificial intelligence programs getting out of hand, similar to the premise behind pretty much any major science fiction film released in the 90s. While several tech leaders, such as SpaceX's Elon Musk and famed theoretical physicist Stephen Hawking, have expressed concern about such a scenario, AI100 researchers said there is really no need to be afraid of AI programs going rogue. The report said that there have been no machines developed with the ability to sustain long-term goals and intent on their own and there are also no plans to create such machines in the near future. The real dangers of artificial intelligence lie not in its potential to become a Skynet-like program, but in the unintended consequences that could come about from a helpful technology such as the displacement of human labor and the erosion of privacy. To avoid such an event, it is crucial for AI researchers and policymakers to find a balance between developing innovations and adhering to social mechanisms to ensure that the benefits of such a technology will be widely distributed.

Benjamin Hiorns is a freelance writer and struggling musician from Kidderminster in the UK who will probably be replaced by a software program at some point in the near future.

Comments

More Leaders

*

Leaders

Inspiring Female Leaders: An Interview with RAPP CEO Gabrielle Ludzker

Gabrielle Ludzker is not just any CEO. The current head honcho at customer experience agency RAPP has spent her career breaking away from the traditional corporate CEO stereotype. and leads to inspire rule breakers. Gabby is an inspirational rule...

Posted by: Benjamin Hiorns
ad: Annual 2024 Now Open For Entries!