How to Outwit AI

Source: Pexels

There’s a global race underway for artificial intelligence (AI) industry dominance. In 1597 English philosopher Sir Francis Bacon wrote in Meditationes Sacrae “ipsa scientia potestas est,” which is Latin for “knowledge itself is power.” However, IQ alone is not sufficient; having a genius IQ does not guarantee success. Similar to human intelligence, having superior capabilities in AI does not necessarily equate to societal success.

Authenticate what is human

What matters is what AI technology is used for and how it is applied. Case-in-point, today AI can impersonate anyone’s voice. Startup Lyrebird claims to have created AI technology that can synthetically mimic a person’s voice from just a 60-second sample [1]. How can someone in know if a voice message or call is really that person? This has potential impact on consumers of voice-recognition products that are in homes, vehicles, and mobile devices. Amazon Alexa, Google Home, and Apple Siri use voice-recognition technology. Authentication will be growth area for enterprising organizations, venture capital, investors, and startups.

Balance AI’s decision-making authority

The global market size for deep learning, a subset of AI-based machine learning, is expected to reach $10.2 billion US by 2025, according to Grand View Research, Inc.’s May 2017 report. Deep learning is a growth area with a major drawback – no one really understands how it reaches its decisions [2]. The quality of deep learning decisions is also impacted by human cognitive bias [3]. The depth and breadth of input data affects deep learning. In any implementation, there should be a manual override to safeguard against undesired outcomes. For example, in 2016 Microsoft’s AI chatbot Tay was shut down after it produced racist and sexist comments within only hours of being live on Twitter [4]. Given these areas of vulnerability, it is important to implement a framework of checks and balances when it comes to how much decision-making autonomy is given to AI systems.

Prepare for societal issues before AI upheaval

By 2025 AI automation will replace 16 percent of U.S. jobs [5]; both white-collar and blue-collar jobs will be eliminated [6]. In a CNN interview on February 15, 2018, British billionaire entrepreneur Richard Branson voiced his opinion that universal basic income (UBI) will one day be necessary due to automation [7]. Universal basic income is the concept of giving every citizen a minimum income regardless of employment status. Opponents of UBI say that a government handout is not the answer; it would disincentivize the workforce. American billionaire businessman Mark Cuban called universal basic income “one of the worst responses” to unemployment caused by artificial intelligence and robotic automation [8].

In addition to unemployment, AI automation will require rethinking education. Business executives and leaders will require a conceptual understanding of AI in order to stay globally competitive. Top-ranked business schools such as Harvard, MIT’s Sloan School of Management, the Kellogg School of Management, and INSEAD have already integrated AI in their MBA curriculums [9]. Displaced workers may require training and learning new skills in alternate professions.

AI can be applied for advancing humanity or nefarious purposes; this is the double-edged sword that applies to any innovative technology. Timing is important. Now is the time to think ahead and prepare in advance for the potential pitfalls—before AI becomes ubiquitous.

“Risk comes from not knowing what you’re doing.” — Warren Buffett

Copyright © 2018 Cami Rosso All rights reserved.