Although artificial intelligence has existed for decades, the face of modern AI has rapidly transformed amidst the ever-evolving digital age. First referred to as knowledge engineering, algorithm-based machine-learning is currently more focused on the development of reasoning and perception than ever. Today, AI has become one of the most talked-about technologies and shows no signs of ceding the limelight.
What is OpenAI?
First introduced to Silicon Valley in 2015 by tech moguls Elon Musk and Sam Altman, OpenAI began as a simple idea motivated by the desire to build artificial general intelligence (AGI) - an AI system that possesses human-like capacity to reason and apply skills to unfamiliar situations.
According to Greg Brockman
, Chief Technology Officer of OpenAI, “The creation of beneficial AGI will be the most important technological development in human history, with the potential to shape the trajectory of humanity.”
Though existing AI systems have proven superior to human intelligence, OpenAI aims to engineer more impressive systems with the capacity to surpass human capability on a larger, more intimidating scale.
No longer are ultra-smart robots mere content for apocalyptic sci-fi movies, they’re tangible modern realities that are being tirelessly developed and employed to enhance the human experience. But how exactly does OpenAI’s mission to “ensure that artificial general intelligence benefits all of humanity” fit into the bigger global picture? How is OpenAI changing our world?
What is the goal of OpenAI?
OpenAI is the for-profit sector of the non-profit organization OpenAI Inc. OpenAI’s goal is to research and build friendly artificial intelligence
. According to Musk, OpenAI aimed to “develop artificial intelligence and apply it to [OpenAI’s] significant goals, such as bringing benefits to all humanity and not being bound by the requirements to generate financial profit."
OpenAI was also founded, in part, as a working solution to Musk and Altman’s growing existential concerns about the potential for large-scale disaster as a result of careless misuse of general purpose AI. The San Francisco-based company has a long-term focus on making serious fundamental advances in AI, grounding their target toward society-friendly technology.
How is OpenAI different?
OpenAI’s approach is unique to the slew of big tech companies venturing into the world of artificial intelligence. The working business model is completely unconstrained by a demand to generate a financial return and supremely driven by the call for advancement in the digital intelligence field.
OpenAI believes that artificial general intelligence should function as an extension to society, broadening the horizons of the human experience. By joining forces with some of the most influential companies and names in tech, they are in good company.
In fact, in July 2019 OpenAI accepted a $1 billion investment from Microsoft. The jaw-dropping deal provided the AI company with more than enough financial artillery to compete with their well-funded AI-driven peers like Chinese tech giants Baidu and Tencent.
Their introduction to the world was incredibly well-received despite the inevitable hesitation toward engineering robots with near-human cognition. The startup has grown an impressive list of investors including PayPal co-founder Peter Thiel, LinkedIn co-founder Reid Hoffman, and more broadly, Google, Amazon, Facebook, and more.
OpenAI: A brief history
December 2015: OpenAI is announced to the public
OpenAI releases an official mission statement
. Altman and Musk are named the key founders, and Ilya Sutskever is revealed as the chief scientist. A cast of world-class scientists, engineers, and researchers were also named, including Trevor Blackwell, Pamela Vagata, Pieter Abbeel, Vicki Cheung, Andrej Karpathy, John Schulman, and Wojciech Zaremba.
August 2016: OpenAI Gym is released
Although the OpenAI Gym was first published as a beta version, the fully-modified and functional open-source library was released several months later. OpenAI Gym was the company’s premiere suite of tools for developing and comparing machine learning algorithms.
November 2016: OpenAI takes on U.S. Congress
Summoned to what was billed as the first congressional hearing on artificial intelligence in American history, OpenAI representatives discussed the future of AGI and how it could prove to be an asset to the nation’s economic development.
August 2017: Dota 2 1v1 bot beats top professionals
Engineers at OpenAI designed an AGI bot that learned the ropes of Dota 2 entirely by means of self-play. The bot was able to defeat the world’s top Dota 2 professional players under standard tournament rules. In a game that requires players to study and predict the actions of other players, OpenAI’s bot served as a significant stepping stone toward building AI systems that accomplish defined goals in complex situations involving real humans.
February 2018: Malicious uses of AI report released
Harkening back to Altman and Musk’s original concerns, OpenAI’s founders co-authored a forecast report on AGI misuse
that explored the global security challenges that the company faces. The report was a collaborative effort among researchers at the Centre for the Study of Existential Risk, the Electronic Frontier Foundation, the Future of Humanity Institute, and the Center for a New American Security.
April 2018: OpenAI Charter released
The OpenAI Charter
detailed the principles the company employed to make progress toward its long-term goals. The document highlighted the strategies used, feedback received, and projected timeline of AGI’s development. The Charter effectively covered the company’s commitments to long-term safety, technical leadership, broadly distributed benefits, and cooperative orientation.
August 2018: OpenAI Five outsmarted top 99.95th percentile Dota players
February 2019: OpenAI announced GPT-2
GPT-2 was OpenAI’s more impressive successor to the first GPT. Functioning as a large transformer-based language model, GPT-2 was designed to predict the next word in 40GB of internet text. GPT-2 proved to be a monumental upgrade from GPT, boasting 10x the amount of data and over 1.5 billion parameters.
March 2019: OpenAI LP announced
Announced as the company’s new capped-profit sub-company, OpenAI LP
set its sights on seeing that employees and investors got a capped return on investment should OpenAI successfully achieve their goals. In turn, OpenAI LP also aimed to raise investment capital and attract employees with equity.
July 2019: OpenAI partnered with Microsoft
Tech titans at Microsoft invested an incredible $1 billion to support OpenAI’s aim toward building artificial general intelligence. OpenAI and Microsoft planned to collaborate on developing new Azure AI supercomputing technology, with Microsoft’s Azure as the primary cloud provider.
What are the potential benefits of OpenAI?
Pop-culture depictions of artificial intelligence showcase the incredible power of AI but typically zero-in on the apocalyptic doom rather than exploring the human-beneficial possibilities
. And when scientific masterminds like Stephen Hawking claim
that “the development of full artificial intelligence could spell the end of the human race,” there’s plenty of reason to feel uneasy as we dive into a future powered by computers.
Modern tech companies engineering the now-very-real technology behind your worst Terminator dreams are making a determined effort to weigh both the benefits and the threats on their pathways toward development. OpenAI is a shining example of these good and evil precautions taken with pointed gravity.
However, when it comes to the many ways human-like robots can propel society forward, enthusiasts agree that centuries of technological advancements could arrive, effectively, all at once. This rapid headway has the potential to transform life as we know it, making serious changes in medicine, green energy, food production, and more elemental facets of human existence.
Let’s take a deeper look at the many ways AGI could begin to impact modern industries.
Though the presence of artificial intelligence in the healthcare field is only in its most nascent stage, the potential for supremely accurate disease diagnosis, drug discovery, patient monitoring, and big data analysis shows incredible promise.
AI in education can effectively eliminate boundaries when it comes to global learning. From simplifying administrative tasks and increasing IT efficiency to adopting personalized learning skills, the introduction of AI into the field can help both students and educators.
Autonomous cars have been a bit of a buzzword for the past few years. Though most recognize Tesla as a primary innovator, many major auto manufacturers are working toward creating self-driving cars. AI-integrated cars have the potential to increase transportation efficiency while reducing the on-road risk of collisions.
The manufacturing industry will likely be among the first to see the power of AGI in action. AI robots can take to assembly lines using visual inspection, predictive maintenance, and analytic sensors to ensure quality is seamless and ceaselessly up to par.
5. Customer Service
Today’s most popular examples of AI come in the form of smart home assistants
. On a larger scale, many tech giants are working to employ artificial intelligence in the customer services fields to provide adequate solutions without the need of any human interference. This could transform how you check out at the grocery store, how you schedule salon appointments, or how you fix malfunctioning devices.
What are the potential risks?
The development of AGI has been often likened to the development of electricity - functioning as not just one simple breakthrough, but as a catalytic enabler for boundless consequential inventions and innovations. However, the power of AGI is one that must be handled with incredible caution and sensitivity as it could quickly transform into an unstoppable harbinger of existential doom.
Though most scientists and researchers agree that a superintelligent AI won’t likely be able to express human emotions or become intentionally good or evil, they do believe that there are very serious extraneous risks that come with the territory.
Let’s dive into the many potential dangers of AGI and how they could threaten humankind.
1. Autonomous weapons
Artificial intelligence programmed to inflict harm are easy recipes for disaster. Researchers have even proposed that the adoption of self-firing weapons could replace the threat of a global nuclear arms race with an autonomous weapons race. Though the concern of the machine gaining a mind of its own is not one technologists are worried about, they are concerned about the devaluation of human life.
2. Invasion of privacy
George Orwell’s 1984 could not have better predicted the current digital circumstances of the Internet Age. Today, it is possible to track and analyze a person’s every action through their online activity. Facial recognition and identity data-basing algorithms exist in a number of everyday formats that effectively calculate who you are.
Take China’s social credit system, for example. This ultra-intelligent technology will be capable of tracking each of China’s 1.4 billion citizens
and delivering a personalized scored based on behavior, both online and offline. Though an interesting idea with plenty of AI-integration, it’s not a far cry from a total invasion of privacy.
3. Social manipulation
More today than ever before, internet users are realizing the power of targeted marketing. In fact, it’s the high-level accuracy that tends to be cause for alarm. Autonomous-power algorithms are employed to understand who you are, what you like, and what you think. However, when abused, these same algorithms can be used to spread misinformation in whatever format will appear most convincing to the user.
Because AGI will be able to collect, learn, and analyze such a vast amount of information about you, there’s reason to believe that those machines could use that information against you. The machines are designed to learn and decide for themselves, but it’s important to remember that the humans behind these futuristic machines carry their own biases.
According to a study completed by the University of Melbourne, AI ‘learns’ by referring to the data that humans ‘feed’ it. If certain groups of people are left out of the data set, the automated process won’t capture their characteristics. This setback alone is inherently the opposite of human-friendly AI.
The future of AGI at OpenAI is one that sparkles with serious global promise. Altering the human condition with the help of human-like robots may be the content of classic movies, but with research and safety-focused companies like OpenAI at the helm of progress, there’s little to fear. OpenAI researchers keep a regularly updated progress tracker
on their website, giving the public access to milestone releases and papers. For all other things tech, you can count on HP Tech Takes.