This is our 9th weekly edition of inspirational reads about Artificial Intelligence for Businesses. Let’s immediately dive into the content for this week:
- The omnipresence of chatbots
- AI translation boosts eBay's sales
- AI resurrects Dali in Florida museum
- How AI is tracking sex traffickers
- The Machine Learning Race Is Really a Data Race
- Legal is a great place to Kick-Start great NLP
FIGURE OF THE WEEK
Chatbots are on the rise, with their 24/7 availability almost every major companies uses them as a customer service tool. It could be used for emergency response, forwarding conversations to the proper employee and placing simple orders.
Juniper Research predicts that chatbots will save companies up to $8 billion annually by 2022.
AI IN PRACTICE
- AI translation boosted eBay sales more than 10 percent by James Vincent for The Verge
Researchers from Washington University found that the introduction of AI powered translation in the US and Spanish speaking countries in Latin America increased Ebay’s sales with 10,9% in 5 years.
Thanks to the implementation of this AI application the language barriers between countries was overcome according to the researchers. For example, Spanish speaking visitors of the website do not have this barrier anymore, because products being described in English are now automatically translated to Spanish.
- AI Has Resurrected Salvador Dali And Now He's Your Museum Tour Guide by Suzanne Rowan Kelleher for Forbes
This article is another example of how a company found a use case for AI to improve customer journey and increase sales. The Salvador Dali Museum in Florida collaborated with an advertising firm to bring Dali back to live as a tour guide for museumgoers. The objective is for visitors to have the explanation about the art works by the painter himself, thus the model AI used interviews, photos, quotes and archival footage in order to be very similar to the real Dali.
After more than 1000 machine learning hours the AI not only knew what Dali sounded like but also how his lips and hands moved when speaking. The AI solution can generate 125 interactive videos and 200 000 combinations. Hence thanks to AI, the museum can provide a different experience to every visitor.
- How Artificial Intelligence Is Tracking Sex Traffickers by Liz Brody for OneZero
Warning ! This is a very long article, but a lot of applications of AI are being explained and you walk through the entire process on how to implement AI. The law enforcement officers do not have the manpower to investigate thousands of photos on the phone of the trafficker and link them with women being placed on escort websites.
This is where Traffic Jam comes into play. This company uses AI to analyse the pictures of the phone and cross reference them with every online website. It can easily identify which picture could be found on which website and when it was posted.
A very interesting article that shows the added value of AI in a completely different field that people are used to.
WHAT TO THINK OF WHEN IMPLEMENTING AI IN YOUR BUSINESS
- AI Challenges And Why Legal Is A Great Place To Kick-Start Great NLP by Mark Sears for Forbes
Legaltech is on the rise, and AI plays a great role in that trend, thanks to Natural Language Processing (NLP). NLP is among the fastest growing applications of AI. But it is also a difficult one. NLP is the application of making a computer analyse and understand human language. Languages have a lot of words or sentences that need context in order to understand the meaning. For example, it will be very difficult for a computer to know when someone is being ironic or not.
The author notes that ironic remarks will not be found in legal documents, so the legal department is a great place for NLP. A model can scan hundreds of pages and legal documents in seconds and give back insights on proximity between documents, and therefore how similar this legal case is to another.
- The Machine Learning Race Is Really a Data Race by Megan Beck and Barry Libert for MITSloan
That data and machine learning go hand in hand is not new, but companies often forget that they have to create accurate data points in order to have an effective machine learning model. The article notes that some companies are racing to apply machine learning to important business decisions, only to realise that the data they need doesn’t even exist yet.
Other companies fail to differentiate themselves as they are addressing a business case with exactly the same data as their competitors. The article concludes with 3 recommendations:
- Differentiated data is key to a successful AI play
- Meaningful data is better than comprehensive data
- What you know should be the starting point