Artificial Intelligence: "Stairway to heaven" or "Highway to hell"
Show an AI-enabled algorithm a photo of a woman and her dog enjoying at the beach, and the computer screen shall read the right results almost every time. The algorithm will also be able to detect a smile on the woman’s face. Yet, for humans, looking at this picture shall evoke emotions and memories. And this is what engineers are trying to infuse in the current prowess of AI. If future algorithms are able to understand advanced human emotions and thought processes, AI could either lead us to the "Stairway to Heaven" or the "Highway to hell".
In a TED talk, I watched recently, when Google scientist Margaret Mitchell showed her computer a series of images depicting a burning home, the machine responded, “This is an amazing view! This is spectacular! ” The computer detected the high contrast and the bright colors and mistook a devastating event for a positive one. Part of the issue is that humans tend to use positive images to describe their experiences, creating gaps in the way AI interprets and understands human reality. Moreover, data – such as photographs – contain biases. Narrowing to a single data set or problem is customary in AI research. But this approach risks developing AI that amplifies human biases. Humans grow and change slowly and thus have time to address their flaws. AI evolves quickly, so immediate correction of any biases is crucial. Technologists must think about how the AI they’re developing will look in five or ten years.
“AI can turn out in many different ways. But in this case, it isn’t a self-driving car without any destination. This is the car that we are driving”.. Margaret Mitchell, Google
Visionaries including Stephen Hawking, Elon Musk and Bill Gates have sounded alarms about the threat AI poses to humanity. Such immense risk calls for an open dialogue about AI. Technologists should share their experiences of computer intelligence and the problems or benefits they foresee. The conversation will raise awareness about where AI is now, where it’s headed and what technologists must do to ensure a beneficial outcome for humanity. Some technologies are obviously helpful. Future AI could, for instance, narrate the world for those with impaired vision. But some technologies magnify harmful biases. Decisions that people make today about AI will shape the future of the technology. Thus, now is the time to clarify plans and goals for AI, such as promoting equality and diversity and aligning it with human culture and the environment.
Meet me - I am Shrey Deepum, a Business analyst and a workplace artist in making. On a normal day, I solve diverse business problems which help me being relevant and future ready. Would love to have a dialogue with you. LinkedIn || Facebook