EticaAI / forum

Etica.AI is concerned about long-term social & ethical implications of artificial intelligence on Africa/Latin America. Do it via grassroots acts in short-term
https://etica.ai
The Unlicense
9 stars 2 forks source link

Implement chatbot (text, voice) interface for public resources of Etica.AI to value human resources #29

Open fititnt opened 7 years ago

fititnt commented 7 years ago

The current reality of Etica.AI is not much different from the future, in the sense that it would be easier to have the machine resources at disposal than human resources. Other groups, private or government of course have more programmers nowadays. Here we can be creative, without reinventing what already exists: after all, a lot of what matters is open-source software in the present day in what Etica.AI proposes, so it is worth treating humans very well so that everyone who has an interest can collaborate.

It's easier for us to create interfaces that help non-specialist humans in data science than to train them to use software. I like the idea of ​​using as reference people of different ages and technological affinities. And I want this not only because I care for web accessibility for many years now, but also because I want people like my grandmother and even my future children to have more powerful interfaces than we have today. And yes, personal motivations make anyone go further, it is not worth it to help people who already work on causes of social impact: it is a win-win situation not only for those involved but for technologies (or teaching methods) which we can slowly popularize in the next few years.

Interfaces for humans (e.g. "do it in a way that your child could use")

By powerful interfaces, "no interface" is a type of interface: use of natural processing language, whether as text or with voice to make people productive, not just release open software or APIs that require programming by human programmers to be useful.

Yes, we can release or code ready to integrate into larger APIs, but among more than one option, I believe it is always worth valuing the human side.

Conceptually speaking, an agent (whether it be a human agent or an artificial intelligence agent) can use a number of specific purpose agents to accomplish a task. This issue here addresses that we should pay attention, even if the code implemented does not look pretty, to appreciate that humans can also access more primary resources. And we should worry about it being accessible to those who need it.

Potential benefit case study: anti fake news

The issue Anti-AI: Fake News repositories #28 started the repository in Portuguese, the Anti-AI: Fake News - Português. At the issue https://github.com/EticaAI/anti-ai-fakenews-pt/issues/2 I made the proposal of, to make a complex problem (automation to know if a news is fake news) divided into smaller steps. And these smaller steps can be weak AIs used by other projects, so that even without a dedicated source of income in the coming years it is plausible to help humans or AIs make decisions about this problem.