tomorrowthink

Can A.I. Tell You a Secret?

Panelists representing a broad spectrum of industries in news/information, hospitality, and technology met at the Dentsu Aegis Network Beach House on Wednesday, June 20th, for a conversation on Artificial Intelligent (A.I.) as it exists today – both its real implications and its limitation – and its future impact on businesses and consumers.

 

Matthew Williams, Head of Creative Impact at YouTube, explained that use of A.I. is shifting the way advertisers approach and engage the consumer to serve timely and more relevant ads, making the user experience more valuable.  Consumers value assisted experience on the front end and that requires rethinking creative.  For example, we are no longer addressing a 24-year-old male, he’s a 24-year-old male who loves camping and is in the market for new car. With personalization, we can avoid assuming all 18-24 year olds are the same.  They are not.

Steve Sirich, GM of Microsoft Advertising, said A.I. and the cloud are now the anchors of their corporate business.  From an ad perspective, A.I. has broadened Microsoft’s agenda with its partners, helping to move the focus from traditional ad tactics to exploring ways to unlock technology and open opportunities for business enterprise and for what he called “technology for good.”

 

He went on to caution with opportunity there is tremendous responsibility.  Microsoft defines responsibility in terms of privacy, security and ethical leadership.  How do we as an industry get behind A.I. in a way that dignifies and demonstrates integrity?  It has to be safe. It cannot harm.  Developers have to be accountable for what they are building.  A.I. must be fair, without prejudice, and inclusive without bias.  It must be transparent so we know when and where it is present, and what it knows and does not know about us.  A.I. is not just Machine Learning and data science.  Human judgment must be a part of the equation.  At its essence, A.I. is about amplifying human ingenuity.

 

Matthew Williams with YouTube put it this way, “Machines provide speed and scale; humans provide the nuance.”

 

Marc Lavallee, Executive Director R&D at The New York Times, shared how A.I. has worked itself into every nook and cranny of the news/information business.  He believes his business has a unique duty to explain A.I. to the world using actual reporting to explain and show what is real today and what is not here yet.  

 

He described a specific application of A.I. to their live coverage of the royal wedding.  Looking for a breakthrough opportunity, the New York Times enhanced their livestream coverage by providing a layer of real-time analysis.  They used facial recognition to identify wedding guests, then took it a step further by adding bios or other additional information that gave viewers expanded insights into the lives of those public personalities.  The intent was not to detract from the wedding or the attendees, but to enhance the viewers’ experience.  In effect, they answered questions the viewers were just beginning to ask.

 

Antoine DuBois, SVP of Global Strategy, ACCOR HOTELS, said A.I. is already playing a major part in hospitality industry pricing, and his focus now is on personalizing the customer experience by anticipating a guest’s needs in order to provide a seamless customer journey.  The biggest challenge is balancing privacy with A.I. personalization.  Most guests will appreciate A.I. working in the background, for example, using facial recognition to help staff personally greet a guest upon his arrival.  It is more difficult, though, to find a good balance between A.I. and privacy in the room when guests are wary of the presence of A.I. devices.

 

The panelists closed in full agreement on one overarching priority as we embrace A.I. and Machine Learning to move into the future.  It is important we to build an ecosystem of interoperability and partnership, where we work together utilizing the best of each.  Combined we can offer so much more to the consumer. 

 

For more, watch the recorded session here: