Given the proliferation of AI (Artificial Intelligence), social scientists try to find new ways of thinking about the society in general while the engineers developing these AI algorithms try to deal with the social and ethical conflicts arising from their outputs. Given these conundrums, there is a need for engineers and social scientists to collaborate to ensure that algorithms do not amplify existing individual biases.
Although major stakeholders within the AI industry- especially those located in Silicon Valley- develop better systems and produce more user data, in order to be able to train their systems more accurately, insights from the social sciences and other fields are rarely taken into consideration.
As machines continue to re-shape major dynamics within the society (such as how we work and communicate with each other) economists and philosophers along with other social scientists need to stay abreast of the latest developments. Given this mutual connectedness, it is crucial that both computing and social science pay attention to each other.
Still, such a mutual collaboration between computing and social science would not guarantee all solutions to our ongoing social and economic problems. As long as there is a questioning of conventional wisdom no technology would alleviate our issues regardless of how sophisticated it might be.
To reach new heights of understanding, both computer and social scientists need to question their conventional wisdom from within.
Amidst the complexity of our world, the best computer scientists can do is to model the world by simplifying it. Yet, the results would differ based on the amount of imperfect information added. In order to model behavior in a particular case, an empirical method needs to be followed. Unlike the natural sciences, AI would progress through a richer set of models that take into account the variety of social experience rather than merely focusing on analytical frameworks.
One of the major weaknesses of computer science is its overemphasis on analytical methods at the expense of breadth regarding the social, economical and historical factors. Computing jobs- especially those in the field of AI- now require applied mathematicians and statisticians rather than a real computer scientist. There is also a need to understand history, sociology, and political science among other disciplines to become a true computer scientist.
Today’s computer scientists did, in general, have not spent much time thinking about social problems or studied other disciplines such as development studies or economics related to social science. The current job market tends to reward those with technical skills rather than interesting research agendas which aim to solve society’s critical issues.
The rules and techniques of the computing profession such as mathematics, the standard optimizing, and the established ML (machine learning) tools simply ensure that the analytical frameworks are empirically coherent. Yes, an excessive focus on these techniques constitutes a problem itself due to being myopic to other crucial factors that might help to resolve the problem.
There are two paths for an individual who would like to succeed as a computer scientist:
- Either to come up with a new technique or piece of evidence to shore up conventional wisdom.
- Or to challenge conventional wisdom.
While the latter might sound as a high risk yet high return strategy, if done correctly, it would certainly pay off.
Computer science should be conceptualized as more than following the rules of thumb. The field of computer science teaches us to think in conditional terms as different remedies might be necessary based on the different constraints. Such a way of thinking results in a contextual type of diagnostic approach rather than a blueprint approach as if set into stone.
The lesson to be learned by younger computer scientists is: Specify an intellectual area that has exceeded what the theory and empirics can support and chip away at it. Yet, do so without abandoning too much the discipline’s accepted methods. In other words, find your own middle way!