As a San Francisco resident, I've often been awed and inspired by the sight of the Golden Gate Bridge. A team from MIT's Media Lab ran a photo of the landmark through its "Nightmare Machine" and now it looks like a moving, tentacled monster that will grab and crush any car that dares to cross it. The Nightmare Machine uses deep-learning algorithms (and possibly evil spirits) to turn ordinary images of people and places into scary ones. To help the AI learn maximum spookiness, the public is invited to rate the faces as "scary" or "not scary." The Halloween-perfect project comes from MIT Media Lab's Scalable Cooperation group, which studies how technology is reshaping the nature of human cooperation.

The world's largest education company is leveraging IBM's Watson platform as it tries to take college tutoring from campus libraries to the virtual world. Pearson is partnering with Armonk, New York-based International Business Machines Corp. to use the Watson artificial intelligence product as an online tutor for college courseware. The companies on Tuesday announced a pilot project that's already underway in the U.S. and is set to expand through 2017 and 2018. Both companies declined to disclose terms, costs or revenue projections from the venture. The project is part of Pearson's efforts to shift its business into the digital age, as it struggles with slumping textbook sales and lower college enrollments in the U.S. IBM is seeking to drive revenue growth by developing practical applications for Watson, its software that wooed the sector five years ago by beating two human champions on the TV game show Jeopardy!

The same internal, deep learning tools that Microsoft engineers used to build its human-like speech recognition engine, as well as consumer products like Skype Translator and Cortana, are now available for public use. Redmond announced today that it is open-sourcing the Cognitive Toolkit that has led to many key developments coming out of its dedicated AI division. In other words: anyone can now train their own artificial intelligence. Formerly known as the CNTK, Microsoft says the beta version of the Cognitive Toolkit is not only faster than previous incarnations, but it is also beats out competing deep learning toolkits – especially when crunching large datasets across multiple machines. On a more practical level for startups and hobbyists, Microsoft says the platform is flexible enough to run on a solo laptop -- just in case you don't have a server farm loaded with NVIDIA GPUs at your disposal.

We approach this problem as a classification problem, predicting two classes -- whether the flight will be delayed, or whether it will be on time. Broadly speaking, in machine learning and statistics, classification is the task of identifying the class or category to which a new observation belongs, on the basis of a training set of data containing observations with known categories. Classification is generally a supervised learning problem. Since this is a binary classification task, there are only two classes. To solve this categorization problem, we will build an experiment using Azure ML Studio.

The next phase is augmentation and modification. As media planners and strategists are freed from mechanical tasks, they can focus on understanding how media mix can inform creative work. Right now, all the creative work is done up front; once a campaign launches, it becomes a matter of optimizing placement and timing. Down the road, machine learning may help recognize when the content itself is the problem, and also campaign workflows that are more responsive to news events, for example stopping a programmatic run to lead on-the-fly creative that resonates with a stunt that just went viral at Burning Man or an October surprise in the political world.

In this guest post, regular B&T contributor and industry veteran Robert Strohfeldt harks back to the good old days of marketing and tells us why artificial intelligence will never replace real-life creative intelligence. After seeing a recent report that Think TV is planning to use Artificial Intelligence for ratings, I tried to think back to when this "digital" madness stated. Only an idiot would deny that technology has added both quality and quantity to our lives. To those who hanker for the "good old days", I will give you one word to make you realise how much better things are in 2016 compared to 1916 – Dentist. But is technology now increasing at a rate which can be overwhelming?

My team and I attended several legaltech-focused conferences this month and of course, as expected, machine learning and AI were the topics that everyone wanted to discuss in between sessions. Yet interestingly, at the Emerging Legal Technology Forum put on by Legalx, one of the panelists -- Mark Tamminga, leader of innovation initiatives at Gowling WLG -- was against using these terms in reference to emerging legal technologies. He pointed out that often what is being called AI is really not that at all, and felt that these words were being used as fancy buzzwords that escape the real mechanics of these technologies. As someone deeply involved in the development community here in Toronto, I wholeheartedly agree with his perspective. Many of the conversations occurring in legaltech around what people are calling machine learning are actually algorithmic solutions preprogrammed (that's right, programmed by humans) to do a particular task; nothing that deviates greatly from anything that's already been done many years ago.

Matt Zeiler grew up in a Canadian farming community -- but fast forward a few decades and he's now running a startup that's looking to bring the same kinds of visual search tools that Pinterest and Google have to other companies and developers. That company is Clarifai, a New York-based startup that offers developers the ability to tag metadata to photos in such a way that the company algorithmically learns what kinds of objects are in photos. With that, Clarifai developers can train algorithms to be able to search for those objects, or input their own photos in order to find similar objects. The company said today that it has raised $30 million. The round was led by Menlo Ventures, with Union Square Ventures, Lux Capital and others participating.

As artificial intelligence (AI) begins to power more technology across industries, it's been truly exciting to see what our community of developers can create with Watson. Developers are inspiring us to advance the technology that is transforming society, and they are the reason why such a wide variety of businesses are bringing cognitive solutions to market. With AI becoming more ubiquitous in the technology we use every day, developers need to continue to sharpen their cognitive computing skills. They are seeking ways to gain a competitive edge in a workforce that increasingly needs professionals who understand how to build AI solutions. It is for this reason that today at World of Watson in Las Vegas we announced with Udacity the introduction of a Nanodegree program that incorporates expertise from IBM Watson and covers the basics of artificial intelligence. The "AI Nanodegree" program will be helpful for those looking to establish a foundational understanding of artificial intelligence.

AI has a long way to go before people can or should worry about turning the world over to machines. Elon Musk's new plan to go all-in on self-driving vehicles puts a lot of faith in the artificial intelligence needed to ensure his Teslas can read and react to different driving situations in real time. AI is doing some impressive things--last week, for example, makers of the AlphaGo computer program reported that their software has learned to navigate the intricate London subway system like a native. Even the White House has jumped on the bandwagon, releasing a report days ago to help prepare the U.S. for a future when machines can think like humans. But AI has a long way to go before people can or should worry about turning the world over to machines, says Oren Etzioni, a computer scientist who has spent the past few decades studying and trying to solve fundamental problems in AI.