For a billionaire looking to leave a lasting impact on the future of humanity, Elon Musk probably couldn’t have chosen a better field than AI research. There is something inherently seductive about the whole premise of designing an Artificial Intelligence. Development of strong AIs would probably revolutionize our society, culture, economy and our very understanding what being human entails. Besides, it wouldn’t exactly hurt Musk if it gets him access to cutting edge technologies that could have far reaching implications on his other pet projects like SpaceX, Solar City and Tesla.
Historically, AI research has had a checkered past forever tied to market interests and commercial applicability. Ever since the first conference of AI researchers in the summer of 1956, funding into the field has followed a consistent “boom bust” cycle. Advancements in the ‘60s and ‘80s led to heavy funding by governments and investors, followed by decades of inertia infamously dubbed as “AI winters”. But it was only with the evolution of computer processing capabilities and the internet age post ‘90s that AI research has managed to find consistent progress and invite massive commercial interest.
The OpenAI project, jointly unveiled by Musk and Sam Altman (of the Y Combinator fame), is a break with current dominant model of AI research that is taking place all over the world. All the tech giants are investing heavily into AI research. In the Silicon Valley, headhunting of AI specialists is reminiscent of sports franchises in the NHL or NBA throwing millions to attract new talent. Google, Facebook, IBM, Microsoft, Salesforce all have invested heavily into AI research involving deep learning, machine learning, data mining, among others.
Academic research and industry requirements are not always aligned. Free unrestricted research can and do get caught up in the race of product cycles and quarterly results. And true innovation in the vast and complex field of AI research is impossible without open unrestricted access to data and new ideas. Open source academic initiatives are essential, as most tech companies are starting to find out. Companies like Google are reluctantly starting to share or release into the public domain at least part of their data on deep neural networks in a bid to foster a free exchange of ideas. Musk’s OpenAI platform takes this to the next level aiming to provide researchers with a non-profit open lab platform to work on cutting edge AI research unencumbered by industry deadlines.
There is another side to the OpenAI initiative. With AI advances coming thick and fast, and billions of dollars being funneled into research by companies as well as governments, many prominent individuals in the science and tech establishment like Bill Gates and Prof. Stephen Hawking have been raising cautionary voices against unbridled and unregulated AI research. We are still in the era of “weak AI” and anything remotely similar to Skynet or HAL 2000 may well be decades away. But AI technologies are still powerful enough to wreak havoc if used with malicious intent.
The only way to guard against such eventualities is to even the playing field by making the knowledge open source and available to all. Research initiatives like OpenAI might very well be the best possible way to ensure that.