OpenAI: The Future of AI

AI has been a proven game changer ever since its founding. Recently, one such game changing innovation with AI has been developed by OpenAI. OpenAI’s mission is to ensure that artificial general intelligence (AGI)—by which the team mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity. The company will attempt to directly build safe and beneficial AGI but will also consider its mission fulfilled if work aids others to achieve this outcome. As a company, OpenAI commit to use any influence obtain over AGI’s deployment to ensure it is used for the benefit of all, and to avoid enabling uses of AI or AGI that harm humanity or unduly concentrate power.

The company trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer follow up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests. ChatGPT is a sibling model to InstructGPT, which is trained to follow an instruction in a prompt and provide a detailed response. “We are excited to introduce ChatGPT to get users’ feedback and learn about its strengths and weaknesses. During the research preview, usage of ChatGPT is free. Try it now at chat.openai.com.”

The company’s AI model called GPT-2 (a successor to GPT), was trained simply to predict the next word in 40GB of Internet text. Due to the concerns about malicious applications of the technology, OpenAI are not releasing the trained model. GPT-2 is a large transformer-based language model with 1.5 billion parameters, trained on a dataset of 8 million web pages. GPT-2 is trained with a simple objective: predict the next word, given all of the previous words within some text. The diversity of the dataset causes this simple goal to contain naturally occurring demonstrations of many tasks across diverse domains. GPT-2 is a direct scale-up of GPT, with more than 10X the parameters and trained on more than 10X the amount of data. As an experiment in responsible disclosure, the team are instead releasing a much smaller model for researchers to experiment with, as well as a technical paper.

The team trained this model using Reinforcement Learning from Human Feedback (RLHF), using the same methods as InstructGPT, but with slight differences in the data collection setup. It trained an initial model using supervised fine-tuning: human AI trainers provided conversations in which they played both sides—the user and an AI assistant. Moreover, the team gave the trainers access to model-written suggestions to help them compose their responses while they mixed this new dialogue dataset with the InstructGPT dataset, which they transformed into a dialogue format.

Reimagining AI Applications
The company has plans to create a more innovative solution for them. There are several things they think are important to do now to prepare for AGI. First, as the team work to create successively more powerful systems, they envision to deploy them and gain experience with operating them in the real world. In fact, OpenAI believe this is the best way to carefully steward AGI into existence—a gradual transition to a world with AGI is better than a sudden one. “We expect powerful AI to make the rate of progress in the world much faster, and we think it’s better to adjust to this incrementally.”

A gradual transition gives people, policymakers, and institutions time to understand what’s happening, personally experience the benefits and downsides of these systems, adapt to the economy, and to put regulation in place. It also allows for society and AI to co-evolve, and for people collectively to figure out what they want while the stakes are relatively low. “We currently believe the best way to successfully navigate AI deployment challenges is with a tight feedback loop of rapid learning and careful iteration. Society will face major questions about what AI systems are allowed to do, how to combat bias, how to deal with job displacement, and more. The optimal decisions will depend on the path the technology takes, and like any new field, most expert predictions have been wrong so far. This makes planning in a vacuum very difficult.”

A Mission to Craft AI-based Solutions
What started off as a research-based organisation, over time, developed an interesting language model that can generate human-like text, and the development of toolkits for comparing reinforcement learning algorithms. In addition to its research and development efforts, OpenAI also engages in policy work, education and outreach, and the development of AI applications. The company today has developed a variety of AI based platforms and made a number of partnerships with in the field of AI, including Microsoft, Google, and IBM.

OpenAI’s founders include Sam Altman, Greg Brockman, Ilya Sutskever, Wojciech Zaremba, and Elon Musk. Sam Altman is the president of Y Combinator and the co-founder of OpenAI. Greg Brockman is the CEO of OpenAI and was previously the president and CTO of Cloudera. Ilya Sutskever is a renowned AI researcher who has worked on deep learning and neural networks. Wojciech Zaremba is a research scientist at OpenAI and a leading expert in natural language processing. Due to its highly innovative capabilities, OpenAI has attracted funding from a number of high-profile investors. Apart from its co-founders, Reid Hoffman, Peter Thiel, and Marc Andreessen were also early investors. “We are committed to doing the research required to make AGI safe, and to driving the broad adoption of such research across the AI community. Moroever, as a company, we are concerned about late-stage AGI development becoming a competitive race without time for adequate safety precautions. Therefore, if a value-aligned, safety-conscious project comes close to building AGI before we do, we commit to stop competing with and start assisting this project. We will work out specifics in case-by-case agreements, but a typical triggering condition might be “a better-than-even chance of success in the next two years.”