AI has been billed as the next big thing in the digital economy, and as is often the case with emerging technologies, there’s plenty of hype and doomsday predictions, in equal measure.
Many emerging technologies have come and gone either because there have arrived too early or are commercially unviable but AI seems different. Automation has always provided huge commercial advantages for businesses, put machine learning into the mix and all of a sudden you have something really powerful that can offer unparalleled competitive advantage.
The question is, how are the big four tech companies – Facebook, Google, Amazon and Apple applying AI to their business models and who’s winning?
Facebook focussed on fighting fake news using AI
As I covered in my recent blog, in the wake of Cambridge Analytica debacle, Facebook has had to double-down on its effort to protect user data, make Facebook a safer environment and fight fake news.
In the recently concluded Facebook F8 conference, Mark Zuckerberg announced that one of the key tools used to fight fake news and bad content on its platforms would be AI, which will identify fake accounts and content based on certain parameters and adapt to changing strategies used by spammers and bad actors on its network.
One of the other ways to stop bad actors in their tracks using machine learning, according to Zuckerberg, is facial recognition technology, which will alert you if someone is trying to impersonate you by using your picture as their profile picture as an example. Privacy campaigners, on the other hand, see this as yet another way Facebook is encroaching in peoples private lives and there is an ongoing class action lawsuit in USA against Facebook on use of this technology.
It seems that due to a number of these difficult issues Facebook is currently grappling with, AI will remain one of the key tools in its fight to eliminate bad content and actors from its network.
Google invests heavily in AI infrastructure
Google has a history of launching services that get used by millions if not billions of people and deep cash reserves allow it to enter markets by usually offering its services for free and worrying about monetising it later.
AI is no different, Google is setting up AI focussed computational architecture across the globe using machine learning Tensor Processing Units but now they have launched TPU 3 which is so powerful that it has to be liquid cooled. This computing power is now available to its cloud customers via Google AI.
In the recently concluded Google IO developers conference, Google CEO Sundar Pichai announced that they are using AI in the healthcare sector and have successfully tested machine learning for diagnosing diabetic retinal retinopathy. The technology went even further in that by scanning the eye for retinal issues, AI also picked up risks of adverse cardiovascular events which may have been completely missed otherwise by a trained medical professional.
There were also examples of using AI to improve accessibility, enhancing photos (turning your old black and white photos to color or vice-a-versa) but one thing that was truly awe-inspiring at Google IO 2018 was human-like Google Assistant with natural conversation capabilities.
Here is a segment from Sundar’s keynote so that you can experience this for yourself:
Wasn’t that truly mindblowing?
Mindblowing it may be but it has also got a lot of people concerned about where AI is going and how machines can convincingly mimic attributes that have previously been quintessentially human.
AI is also being used in advertising, Google’s biggest revenue generator. Google’s native advertising platforms such as Google AdWords and DoubleClick are now more sophisticated than ever and use deep machine learning to optimise campaigns in real time.
Amazon may be the AI darkhorse
Amazon has been using machine learning perhaps longer than most with its recommendations feature (people that bought this also bought..) for years now, notwithstanding the automation and AI used in its colossal warehouses.
Alexa, however, was a real game-changer for Amazon, as the breakaway smart home speakers that could book you an Uber without lifting a finger. Since then Amazon has built capabilities around image recognition, video analysis, chatbots, and AI-driven language services.
Amazon has also packaged it’s deep learning capabilities into its cloud services infrastructure – AWS (Amazon Web Services).
Amazon does not hold it’s annual conference with the same fanfare as Google or Facebook and they tend to be lowkey localised events. Amazon also has a reputation for being secretive about their tech development but there is no doubt that Amazon has its sights firmly set on integrating AI into every aspect of its business model.
Apple turns from AI leader to follower
Apple announced the launch of Siri in 2011 and although not perfect, it reignited the AI race and demonstrated real-world application of technology in the increasingly mobile world.
Since the glory days of Siri, competition has become a lot more fierce and voice assistants such as Alexa and Google Assistant are much better at voice recognition and understanding complex voice commands.
In the recently concluded Apple’s developer conference (WWDC), Apple announced how they are revamping Siri by adding voice shortcuts that add a special meaning to a certain voice command, for example, you can set your thermostat to increase your house temperature with the voice command “I am heading home”.
There is no doubt Apple has a lot of catching up to do in the AI space and their application seems limited to voice.
So who is winning the AI race?
Considering everything that is in the public domain, our money is on Google. They are performing solidly on voice, image recognition, image enhancement, AI led computing, chatbots and integrating AI into most of their product offerings.