By all accounts, Elon Musk is a real world Bruce Wayne in many ways. He certainly has the financial backing for such a high profile gig. He has his own car and a number of tech companies that even reach beyond the confines of our planet. But instead of fighting crime as a caped crusader, Musk is using his resources for a slightly different cause. He’s using his wealth and knowledge for improving our lives with advanced technology. Amidst all of the tangible products and advancements like Tesla cars and Falcon rockets, one project, in particular, stands out as a potential game-changer in the foreseeable future – Open AI.
OpenAI is a non-profit artificial intelligence research company whose mission is to build and promote safe AI as a highly beneficial tool for the human race, rather than seeing it as a threat. The company is a collaboration between Musk, Sam Altman, president of tech incubator Y Combinator and other Silicon Valley entrepreneurs.
OpenAI Vision
What’s most intriguing in the whole story is the company’s approach to this endeavor. In the market of machine learning technology, where the most important thing is presenting the next product or increasing the profit margin, OpenAI is practically giving away the technology for free. Instead of focusing on earnings, the company is offering unrestricted access to their research to anyone interested in it. What might be the most disruptive technology of our time is used solely as an incentive for a better future.
The motivation behind OpenAI is the concern that artificial intelligence presents a latent existential threat for humankind. Musk is quite vocal on this issue, while the co-chairman Altman believes AI will soon surpass human intelligence. As such, almost complete openness was essential in order to achieve results that would be beneficial to mankind. The company wants to make AI better and believes a joint effort, not keeping it to yourself, is key.
OpenAI is absolutely correct in reasoning that AI holds a high level of risk and danger without human control. The $1 billion organization collaborates with companies such as Microsoft and Amazon to research and deploy new technologies. However, companies like the aforementioned duo may use the opportunity to utilize open-source software and data to challenge and eventually level the playing field against companies like Google and Facebook, two tech titans that possess huge amounts of data. This brings us to the main question at hand – how exactly will OpenAI be of use to us, the consumers?
What is OpenAI used for?
Last December, OpenAI released Universe, a software platform for measuring and training general intelligence of AI. It’s applicable to games, websites and other applications by allowing a single AI agent to use the computer as we do. Universe trains an AI agent to execute any task a human can complete with a computer. In April last year, OpenAI launched Gym, a toolkit for developing and comparing reinforcement learning algorithms. Thanks to Universe, any program can now transform into a Gym environment. The December release consisted of various games, ranging from bonafide blockbusters like GTA V to simple flash games to old Atari classics, as well as numerous browser tasks.
The practical use means, for starters, that we’ll get better AI in games. From a strictly consumer-centric approach, the multi-billion video game industry presents a perfect setting for such an undertaking. It’s also a logical step as reinforcement learning is a part of machine learning that examines how an agent can learn to achieve goals in a complex, uncertain environment. To put it simply – it’s based on the idea of a reward for efficient problem-solving. As everyone who touched a controller knows, video games are abundant with such processes. In addition, they have a low level of impact, which makes them a good starting point for more advanced projects.
Unlike video games, other tasks like web browsing are not quite as straightforward. Universe includes a number of browser-based environments which require AI agents to read, navigate, and use the web just like we do. Open AI created “Mini World of Bits”, a way of benchmarking both simple and complex browser tasks. In the beginning, the agents get acquainted with regular interface elements like buttons, sliders, and lists. Then, they perform various basic tasks like clicking a specific button and subsequently move on to more difficult tasks. The company believes its AI agents could perform a variety of extremely complex tasks. Looking up things they don’t know on the internet, managing your email or calendar, booking flights and numerous others are all within AI’s grasp, according to OpenAI.
What is Open AI Universe?
What makes Universe so appealing is that it can easily be translated into real world performance. If artificial intelligence can achieve high levels of accuracy with these kinds of tasks, it will be able to cope with any commands that are currently far beyond the grasp of the best AI products we have today such as Alexa, Google Assistant, and others. That is where the true applicability of Open AI lies in terms of the consumer market. Remember all the times modern technology, especially your personal assistant frustrated you? The lack of understanding the context of your commands, absence of support for specific services, the inability to perform certain tasks and so forth – all that will be a thing of the past.
We’ll get a more encompassing and precise personal assistance and a larger ecosystem of integrated AI products that we’ll be able to interact with. We’ll get better, more intuitive games of both leisure and educational variety, as well as a greater integration of websites and applications for easier use. This is only an imminent, small part of what AI will be able to do – the possibilities are practically endless.
Conclusion
There is no denying that Open AI is a noble cause at heart. As a nonprofit, the company has vowed to share its knowledge and join forces with others to promote the open-source nature of artificial intelligence and push it as far as possible. There is a rationale for concern, though. If you can build AI to do good things, that it stands to reason it can also do bad things. Founders of Open AI are not alone in voicing these concerns. However, they are ones of only few who are actively countering them by domesticating AI. Musk, Altman, and others firmly believe the right way to ensure obedience and usefulness of artificial intelligence is to increase access to it, rather than limiting or restricting it. That is why the idea itself has attracted so many bright minds willing to take part in it.
For better or for worse, Open AI’s dream at the moment is that Universe can drive the development of machines with general intelligence, adaptable and flexible enough to complete other tasks alike. Artificial general intelligence or generalization presents one of the biggest obstacles to machine reaching or surpassing human-level intelligence. OpenAI wants to create the technology that will build a machine that can perform in a way identical to humans. The world could certainly benefit from creations that can break down real, quantifiable data and help with any situation necessary. Nevertheless, the recent boom of AI-powered smart machines has come up short in those terms. It’s a problem that only machine intelligence can successfully tackle and resolve. There’s still some time before that happens, but it is slowly getting there.