A Lack of Focus in My Approach to Writing About AI
A Short Reflection on Different Approaches to Writing
Lately I have been writing exams and keeping up with the daily articles. Writing a little bit every day is a great routine to ensure you at least read some information on a topic, but is it focused? When I posted for suggestions on an AI forum a member made the point about quality as opposed to quantity and I agree. Yet wanting to have a perspective on AI that reaches into a lot of different AI is fun to, and it helps to connect a few dots, so in another sense I am torn. Recurring topics in my writing about AI has been safety, climate, politics, security, techniques, startups and investments. As such it covers a wide range of different areas, yet is not focused or one or the other. What could I gain from being more focused?
Continuity Can Build Towards a Perceived Value
Being self critical for a second there is no guarantee of writing a new article every day for 500 days will add any value. Personally I have been approached by people who have said it has been interesting, yet it is a question of what value a contribution has in the world. What thoughts are valuable and how is it measured? What is good? I am not sure if moralising will do much good at all, and at times that seems futile.
Ben Green wrote a paper called: “Good” isn’t good enough.
In his conclusion he says:
- This requires, first and foremost, a political orientation for algorithmic practice. Rather than referring to “social good,” computer scientists should more explicitly consider and articulate the normative commitments behind their work (whatever they may be).
- Second, in order to actualize these commitments, computer science needs a praxis that engages contextually with the relationship between technological interventions and social impact in both the short and long term. This requires looking to the lessons forged and debated by generations of social thinkers and activists regarding how to actually achieve positive social change. Such reasoning can help computer scientists consider the role of algorithms in improving society, how algorithms can generate unintended impacts when they interact with the social and political world, and when other forms of political action are necessary in conjunction with or instead of algorithms.
- Third, rather than presuming that algorithms provide an appropriate solution for every problem, the field must evaluate algorithmic interventions against alternative reforms. This also means finding new types of algorithmic interventions that better align with long-term pathways of social change. Many of these imperatives draw on the expertise of other fields, necessitating the need for an algorithmic practice that is interdisciplinary at its core, no longer prioritizing technical considerations (such as accuracy) as superior to or more essential than other forms of knowledge.
Why Does this Relate to my Focus?
Likely because I read it yesterday after I saw Ben post it on Twitter. However it is also a way to sum up my focus. My focus has been on (1) political orientation, (2) praxis and (3) long-term pathways for social change. What I find fascinating is that the words ‘climate’, ‘environmental’ and ‘sustainable’ are not mentioned in this context. Due to the wide overarching effect it has upon poverty and inequality you would think this was included. Yet there is my crux: looking at every study and searching for my own focus/interest.
There was however a lot more focus on climate change at NeurIPS, one of the most prominent conferences on AI in the world. You can check out this link.
Partnership on AI had some interesting events on too:
Therefore if my approach is the climate crisis and AI from an anthropological perspective with a keen interest for social data science (programming techniques and theory in social science). Then again that is a wide field, therefore I currently focus on AI, Climate Crisis and Security/Safety. Within this area studying the programmers and understanding the practice is important for me, while keeping an eye on the developments around the world within AI policy and ethics.
Closing in on day 200.
This is #500daysofAI and you are reading article 195. I write one new article about or related to artificial intelligence every day for 500 days.
