OpenAI is giving away $1 million to find out how people want AI to behave.
AI is becoming more powerful and ubiquitous every day. It can chat with us, generate images, write texts, play games, and do many other things that used to be exclusive to humans.
But how do we make sure that AI is doing what we want it to do? How do we ensure that AI is fair, ethical, safe, and beneficial for everyone?
These are some of the questions that OpenAI, a nonprofit research organization that aims to create artificial general intelligence (AGI), wants to answer.
OpenAI has recently launched a program to award ten $100,000 (Rs. 8.25 crores) grants to fund experiments in setting up a democratic process for deciding what rules AI systems should follow, within the bounds defined by the law.
What is a democratic process?
By “democratic process”, OpenAI means a process in which a broadly representative group of people exchange opinions, engage in deliberative discussions, and ultimately decide on an outcome via a transparent decision-making process.
There are many ways such a process could be structured — OpenAI encourages applicants to be innovative, building off known methodologies, and coming up with wholly new approaches.
Some examples of possible questions that a democratic process could address are:
Should AI criticize public figures?
How should AI handle personal data?
What should AI consider the “median individual” in the world?
How should AI balance competing values and interests?
How should AI deal with uncertainty and ambiguity?
Why a democratic process?
OpenAI believes that a democratic process is the best way to ensure that AI systems are aligned with the values and preferences of the people who use them and are affected by them.
OpenAI also believes that a democratic process can help foster trust, accountability, and legitimacy for AI systems, as well as promote social learning and collective intelligence among humans.
OpenAI acknowledges that there are many challenges and limitations to implementing a democratic process for AI governance, such as:
How to define and measure the representativeness and diversity of participants?
How to design and facilitate effective deliberation and decision-making?
How to deal with conflicts and trade-offs among different stakeholders?
How to ensure compliance and enforcement of the outcomes?
How to scale up and generalize the results?
However, OpenAI hopes that by funding experiments in this area, it can learn from the successes and failures of different approaches, and contribute to the development of best practices and standards for AI governance.
How to apply?
If you are interested in applying for a grant from OpenAI, you need to submit a proposal by June 30th, 2023. The proposal should include:
A description of the problem or question you want to address with a democratic process.
A description of the methodology you plan to use for setting up and running the democratic process.
A description of the expected outcomes and impacts of the democratic process.
A description of the team members and their qualifications.
A budget and timeline for the project.
You can find more details and instructions on how to apply on OpenAI’s website: https://openai.com/blog/democratic-inputs-to-ai/
OpenAI will announce the recipients of the grants by August 31st, 2023. The projects are expected to run for six months, starting from September 1st, 2023.
OpenAI will also provide mentorship and support for the grantees throughout the project duration. OpenAI will also share the results and learnings from the projects with the public.
Are you ready to shape the future of AI?
If you have an idea for how to set up a democratic process for deciding what rules AI should follow, don’t miss this opportunity to apply for a grant from OpenAI.
You could be part of a groundbreaking experiment that could influence how AI is governed and used in the future.
You could also help make AI more aligned with human values and preferences, and more beneficial for everyone.
Apply now and make your voice heard!