Cross-Fertilization of Human-Systems Integration and Artificial Intelligence
Human-Systems Integration (HSI) has been defined as a combination of Human-Centered Design (HCD) and Systems Engineering (SE), and an extension of Human-Factors and Ergonomics (HFE) and Human-Computer Interaction (HCI) (Boy & Narkevicius, 2013). What topics do we have within the AI community? The 2020 AAAI conference proposes the following traditional topics: search; planning; knowledge representation; reasoning; natural language processing; robotics and perception; multiagent systems; statistical learning; and deep learning. Summarizing, current AI could be categorized into two fields: data science and robotics. John McCarthy and Marvin Minsky presented the first artificial intelligence program at the Dartmouth Summer Research Project on Artificial Intelligence in 1956 . AI became big during the 1980s but did not go through until recently, leaving the field to HCI during the 1990s and 2000s. We can say that HCI highly contributed to human-centered automation. It is hoped that AI of the 2020s will contribute to the development of human-centered autonomy, which should greatly provide people and machines with appropriate technological flexibility. During the nineteen eighties, artificial intelligence (AI) developed so much that we were thinking that it could invade our lives and replace people. Almost three decades followed with an AI winter! Today, AI resurrects even bigger that before. Should we be worried about being replaced by machines? Or, should we think in terms of interacting and collaborating with smart machines? The Cloud, for example, brings more autonomy to people than any tools had provided before. However, we need to be cautious. We should look after AI algorithms maturity. We should make sure that AI does not bring ways of doing things that are more complicated. Think about voice menus when you call a large company; you usually end up in being extremely frustrated just because the system is too rigid. This is because such voice recognition systems were immature for a long time.
Guy André Boy -CentraleSupélec & ESTIA institute of Technology