Emerging Media Industry Analysis #1

Immerse on 2018-01-01

Artificial Intelligence and Machine Learning

by Kamal Sinclair

Media innovation does not happen in a silo. It happens in concert with emerging technology, advances in science, changes in the cultural sector, and thought leadership from many other fields of knowledge. Therefore, it is important to understand the suite of new technologies and innovations that are changing the human communication architecture and our social systems.

What are the media-related emerging technologies? How will they be used in society, industry and commerce? What is their economic value? Each month for the Making a New Reality project we will look at a different category of technologies and share some of what we’re finding out about them. This month we’ll take a look at artificial intelligence and machine learning.

What is artificial intelligence?

Artificial intelligence (AI) includes a set of linked technologies that enable computer programs to mimic human learning, perception and decision-making. For the purposes of this article, cognitive computing, machine intelligence, machine learning, deep learning, and augmented intelligence are all under the umbrella of AI — although, as you can see in the image below, there is some debate about these categories:

People as the prototype: Deep neural or capsule networks are algorithms loosely based on the networks of neurons in the human brain that are increasingly learning to mimic the intelligence of the human brain. Organizations such as Open AI are contributing to breakthroughs in reinforcement learning, which allows AI to evolve over time through experience. Robots are now learning not only through code, but through visual and aural cues — with new senses being added. For example, researchers are also creating robots that can feel force and are gaining the sense of touch. Companies such as Google are not only copying the brain’s neural networks in their AI design, but expanding to emulate the hippocampus to achieve memory.

Conversational bots take off: AI is being modeled not only on individual human perceptions, but on social interactions. One area of focus of the UX/UI community involves increasing the integration of chatbots or simple AIs into the universe of mobile and voice-driven apps, which can connect the dots between your personal data and apps. This makes navigating functionality on mobile devices extremely fluid for users — but it may also make us more vulnerable to biased algorithms and to privacy issues.

Reshaping industries: Another major innovation area for AI researchers and designers is refining natural language processing and conversational commerce (projected to become a $13.4 billion industry by 2020 ), especially instant translation that will break down spoken and written language barriers between people and cultures. Also, healthcare applications are a focus for increased innovation (i.e., skin cancer detection and eldercare) — even doctors aren’t safe. Finally, modern urban planning will be transformed by AI technologies in the near future, from architecture and infrastructure, to how cities change over time. Imagine these capabilities paired with whatever developers create with the 100,000 questions and answers that Microsoft released to help developers make systems that can read and answer questions as precisely as a human.

A race to outsmart us: The Texas Hold ’em and AlphaGo wins in 2016 were huge milestones for AI advancement, but how long will it really take for computers to match or supersede the human brain? According to Nick Bostrom, an Oxford “actuarial philosopher,” the probabilities of when human-level AI will be attained are 10 percent by 2022, 50 percent by 2040, and 90 percent by 2075. Then again, futurist Ray Kurzweil appeared at SXSW this year and disrupted all the tempered messaging by announcing 2029 as the year AI would exceed human intelligence and asserted that hybrid humans will be the next step.

Whether you believe Kurzweil and Bostrom or not, we are already ten years ahead of schedule. AI is already powering virtual assistants, the internet balloon Project Loon, Google’s DeepMind Wavenet natural language processing, the analyzation of crime patterns, the predictions of weather and financial crisis, and the evaluation of the freshness of strawberries. We are at the very beginning of the adoption of driverless trucks and cars, the ability to pay with your face in consumer transactions, visual search engines, social media by bots, automated wildlife park management, and more.

Hitting the limits: Even art may be up for grabs. Generative art is already being made by smart algorithms and machines (see AI Rapper, the After Us exhibit, IBM’s Watson Gaudí -inspired sculpture). However, experts warn against believing all the hype when creativity and “small data” nuanced perceptions are still the exclusive domain of humans. And with cashier-less stores failing for Amazon because the code couldn’t handle more than 20 customers at a time, we may still be a long way from a full-scale robot invasion.

Learning machines and society

Analysts express very polarized views about how to integrate artificial intelligence into our society in more robust ways. Some futurists are excited by what they imagine humanity will do with their time, once AI frees us from the burden of 47 percent of our current labor requirements, enhances our biological perception to allow higher level understandings of reality, or allows us to shift value systems to better invest in art, social spaces, health, science, or spiritual practice. AI might even help us become “super humans,” according to Chatgrape CEO Felix Hauser, by freeing us from dull, repetitive work.

While the enthusiasts see the incredible opportunities to advance civilization, critics are fearful of the possibilities of unethical and inhumane uses of the new technology or even of AI becoming dominant and controlling in human infrastructure. The recent expansion has caused many observers, policymakers, data scientists and tech gurus to raise red flags about AI that urge companies and regulators to find a humane and balanced approach to the integration of smart machines into our society. For example, Bill Gates called for a tax on robots to offset the severe job loss (potentially a third of US and UK jobs) coming from disruptive new AI capabilities.

Other approaches to managing societal disruption include a focus on augmentation over automation to mitigate job loss. For example, Google Brain is looking to artists to play an exemplary role in this model by adopting the platform Magenta to collaborate with humans to make art. Elon Musk is skeptical that this pairing of human and AI/robot will ultimately be enough to mitigate the power of smart machines. Therefore, he launched Neuralink, a company aiming to physiologically integrate of humans and robots by way of implanting tiny electrodes in human brains.

Finally, some observers are trying to level-set public anxiety about artificial intelligence. For example, Dr. Michio Kaku’s video on what jobs will not get disrupted by machine learning went viral on social media. He draws attention to uniquely human qualities such asimagination, creativity, and problem solving skills that could not be replicated by machines. Demis Hassabis from Google DeepMind also redirects our attention to the great good that can come of AI, and Gary Marcus has been gaining visibility for his TED Talk delineating what current smart tech is good at and what it is not good at.

According to the new Robot Fear Index the message of Kaku, Hassabis, and Marcus may be penetrating the discourse and calming their audiences’ anxieties.

Also, we are seeing a backlash on the bots or smart algorithms that create unfair competition for humans trying to participate in simple processes such as purchasing tickets for a live event. Companies are having to redesign processes to allow humans a chance at participating. Perhaps these kinds of design choices will help to mitigate fears of the robot takeover.

A booming marketplace

There are widely varying assessments of how lucrative AI technology will be — and it’s also worth noting that a number of analysts are warning of a bubble. For example:

Regionally, China enjoys an early mover advantage in AI, unlike other innovation sectors where they are lagging behind the United States (i.e., ride-sharing and financial technology). The government announced plans to invest $15 billion in AI in 2018. Even Google is expanding its artificial intelligence, or AI research to China. However, Canada is planning to give the US and China some serious competition in the area of artificial intelligence. The government has committed $125 Million dollars to launch the Pan-Canadian Artificial Intelligence Strategy. India, will also be a significant player, with an estimated 70% of its companies predicted to integrate AI into their infrastructure in the next few years.

It may feel very challenging for the average person to access and integrate these technologies into their small businesses and personal lives. However, it is increasingly essential for success. Inside Big Data reports that there is a “clear link between an organization’s revenue growth and its AI maturity.” A related poll of 1600 senior executives of large corporations across the globe showed that “organizations who report[ed] faster growth in revenue over the past three years were also more likely to be further ahead when it comes to AI maturity,” indicating that “AI was fundamental to the success.” In fact, 58 percent of US business executives are already using AI and, by 2020, a projected 85 percent of customer interactions will be managed without a human. Will the little people get left behind?

Well, Steve Rosenbush reported in the Wall Street Journal that “the democratization of AI [will] begin in earnest in 2018, at which point our world well may be on the way toward…an algorithm-based civilization.” For example, Alibaba Cloud’s PAI 2.0 (a platform designed to facilitate the deployment of “large-scale data mining and modeling,” with a specific focus on artificial intelligence and machine learning) is supposed to be a game-changer in the democratization of AI to businesses and organizations. Hopefully access to democratized AI will come in time to counter the perpetual and exponential competitive advantages the current players are positioned to enjoy as this industry develops.

The Making a New Reality research project is authored by Kamal Sinclair with support from the Ford Foundation JustFilms program and supplemental support from the Sundance Institute. Learn more about the goals and methods of this research, who produced it, and the interviewees whose insights inform the analysis.

Immerse is an initiative of Tribeca Film Institute, MIT Open DocLab and The Fledgling Fund. Learn more about our vision for the project here.