For many, Artificial Intelligence (AI) is a new concept when in fact the reality is that businesses have been implementing the concept for years. As AI continues to develop and expand across different industries, it’s starting to weave into our everyday lives both at work and at home. From driverless technology to voice recognition “intelligent assistants”, AI technology is seeing major advancements.
In 2016, the concept of driverless cars finally became a reality, with Tesla’s new models being equipped with self-driving technology and Uber’s self-driving cars being introduced in Pittsburg. Not only that, but Google’s artificial intelligence program AlphaGo made history by beating the world champion of the popular Go board game – 10 years ahead of predictions!
However, as artificial intelligence continues to advance, concerns over potential negative impacts are beginning to rise.
Moral and ethical implications of AI
Across the board, experts worry about the ethical implications of AI and for good reason. It’s possible that AI can make mistakes, from reinforcing biases to making racial slurs to not preventing fatal accidents, as in the case of self-driving technology, for example. Where AI interact and deal with human affairs, they will need to be taught what action to take to deal with societal and moral concerns – something that’s not always that straightforward even for humans. The issue of deciding how to ensure AI operate with the same values as humans will continue to be raised as an issue in the future.
Rogue AI and the potential for hacking
While we may not have AI machines going rogue at the levels shown in the Terminator movies, it is possible for malevolent AI to become a reality if the technology used to create these machines is faulty and opens the door to hacking. Cyber criminals could also very well create AI simply for the purpose of deceiving institutions and individuals and hacking into systems that were once though un-hackable.
Impact of AI on the economy
The economic impact of increased automation is a current hot topic and will be a subject of discussion for years to come. Concerns about AI stealing jobs are nothing new and are continually raised by both individuals and the government. In a June 2016 report, Forrester predicted, “Cognitive technologies such as robots, artificial intelligence (AI), machine learning, and automation will replace 7 percent of US jobs by 2025.” This percentage is contingent upon Forrester’s estimation that 16 percent of U.S. workers will be replaced by AI systems and that AI could create new jobs equal to 9 percent of the workforce by 2025. Ultimately, whether you think AI is good or bad for you will depend on the type of job you have. For example, office support staff could see a decrease in available jobs, while automation specialists and data scientists could see an increase in potential positions.
Potential backlash against artificial intelligence could stem from many directions. As AI technology starts displacing more workers and pressure mounts from the public, it’s possible there could be negative political implications. Scientists themselves have also already expressed concerns about the potential risks associated with artificial intelligence. Then, there’s the fact that enterprises and consumers could become dissatisfied with AI if it doesn’t live up to the hype. This could lead overvalued startups to fail and investments to dry up.
Despite the concerns surrounding artificial intelligence, it has recently seen tremendous developments in many areas.
AlphaGo’s historic win again Korea’s Lee Sedol—one of the world’s best Go players—in a landslide 4-1 victory in March 2016 was groundbreaking for both the field of AI and the technique known as reinforcement learning.
Inspired by behavioral psychology, reinforcement learning is an area of machine learning that allows software agents to automatically determine the best action to take within a specific context to maximize performance or the accumulated reward over time. It’s essentially trial-and-error learning as agents learn from the consequences of their actions, rather than from being explicitly taught or even being given overt examples.
Through persistent experimentation and analysis of previous games, AlphaGo taught itself to play Go at an expert level. After beating Lee Sedol, the AI has gone on to win more than 50 straight games against the world’s top Go players as of January 2017. This gives AI researchers hope that reinforcement learning can prove useful in solving real-world problems, such as automated driving. Google has even used reinforcement learning to make its data centers more efficient.
One of the biggest hopes researchers have for AI is that the same techniques that have proved successful in voice and image recognition, among other areas, can prove useful in helping computers analyze and generate language more efficiently. Better language understanding by AI would make machines a lot more beneficial, but the challenge is not an easy one given the complexity of language.
While we are not quite at the level of being able to have deep, meaningful conversations with AI, some impressive strides have been made by Amazon’s Alexa and Google Assistant, the AI virtual assistants taking on Apple’s Siri and Microsoft’s Cortana.
A 2016 survey by TechEmergence found that 37% of the AI executives and startup founders surveyed picked virtual agents and chatbots as their top choice of AI applications likely to take off in the next five years. For their part, companies like IBM and Facebook have been essential to providing developers with platforms to create chatbots. According to IBM, 65 percent of millennials actually prefer to interact with bots than talk to a live agent.
So far, the focus of chatbots has been to understand what the end-user is asking the software and once the question is understood, for the chatbot to either provide a predefined response or do an internet search. In order to make chatbots a must-have for businesses, however, they need to be able to reason, ask intelligent questions and provide tailored responses to end users.
Advancements in AI software and hardware
In addition to creating chatbots, many enterprises are incorporating artificial intelligence into their web apps, mobile apps, and internal business applications. AI is being used to schedule meetings, analyze big data, and make recommendations based on user activity. According to technology research company Gartner, “most of the world’s largest 200 companies [will] exploit intelligent apps and utilize the full toolkit of big data and analytics tools to refine their offers and improve customer experience” by 2018
The AI hardware industry is also experiencing significant advances. In an effort to make AI “available to all”, Intel recently announced its intentions to build AI capabilities into its chips. Other developers are working on AI-powered robots, drones, and autonomous vehicles, with AI hardware revenues expected to increase at a compound annual growth rate (CAGR) of more than 60 percent over the next five years, according to the International Data Corporation (IDC).
AI startups and the impact of tech giants
The increasing interest in AI has led to the founding of many new AI startups. In its most recent report on AI startups, Venture Scanner identified 1,500 AI startups from 73 different countries with more than $9.1 billion in total funding. However, while machine learning has generated major interest, a wide gap still exists between research and beneficial use cases in the real world.
Many companies simply don’t have the resources to drive AI innovation and deliver it at mass scale. That’s where today’s leading technology companies come in. Recently, many large tech companies like Google, Apple, Facebook, Microsoft, and Intel have gone on a buying spree, snapping up a lot of small AI startups and showing no signs of slowing down. With growing investments from enterprises, AI is expected to generate improvements in products and services, as well as the customer experience.