February 22, 2019
Elon Musk is best known for standing at the forefront of IOT innovations such as driverless cars, and outer-space phenomenons which many find mind-blowing even at the current conceptual stage. However his most recent venture is one that has gone above and beyond that of which anyone thought possible at this point in time. Partnering with influential startup backer Sam Altman in 2015, the Tesla co-founder and CEO set up a non-profit organisation, appropriately named OpenAI.
The platform created with the backing of such influential figures, was founded with a mission to reinforce the importance of research and knowledge moving forward into the ever-changing technological revolution. Labelled as the most significant technology to have ever been created by humans, Artificial General Intelligence (AGI) is warned to drastically change the world as we know it. With that, comes the motivation to build a safe environment in which research can be thoroughly tested and explored. This is in order to ensure the benefits of this technology are prioritised above the fear of a potential negative impact. Focusing on long-term research has its’ benefits, however the most recent findings have notably sinister contrast from previous projects.
GPT-2, the enhanced version of its predecessor GPT, is the text-generating system which can write page-long responses with as little as a prompt. The machine can mimic an alarming range of styles and tones from sci-fi prose to fake celebrity and political news, even branching out to educational assignments. Whilst as a concept, the system deserves huge recognition and has undeniably captured the fascination of the globe, the reality doesn’t paint such an idealistic picture. Due to the fear of potential misuse and harmful ramifications as a result of its public release, creators have been pushed to make an unusual decision. The majority of research projects undertaken by OpenAI are ultimately available for public viewing, however it has been decided that this research is to be with-held for the moment, in order to allow more time to discuss the consequences of this immense breakthrough.
Many may jump to the misguided conclusion that this is as a result of insufficient findings or a lack of results, however this could not be further from the issue. It is said that the revolutionary AI system is so advanced that the risk of malicious use is well beyond that of which we are prepared to feasibly control. The initial goal was for it to merely predict the next word in a sentence by assessing the previous content. After a training period of trawling 8 million web pages, it was clear that the results were hugely underestimated. Despite the entirety of research details not yet live, the organizations blog pages are regularly updated with contextual results, most recently an essay assignment written entirely by an AI powered text generator with the only human-written input being a single sentence prompt.
Depending on the environment of which this system is used on its release, it has the potential to massively diversify any number of processes. While the GPT-2 may not be the first of its kind, it is by far the most advanced. With the ability to skew reality in your favour one way or another, it is essential not to be naïve when considering the dangers of artificial intelligence. However fearing its progression will only temporarily stall its inevitable rise. With every new technology that is introduced, there is a period of scepticism from the public, but with time it is subtly integrated into our everyday lives. Elon Musk is just one of many presenting opportunities to grow and evolve as a digital community, and with trust and correct implementation, we will have the ability to harness innovation and take beneficial risks.
“When something is important enough, you do it. Even if the odds are not in your favour.”
– Elon Musk, OpenAI Co-Founder