Media outfits from respected broadsheets to renowned technology websites recently reported a breaking news about the Elon Musk-funded artificial intelligence (AI) software that has the ability to build an authentic-looking fake news story with minimal resource in a matter of seconds. With the current global problem on the spread of disinformation, including those for political propaganda, cultural attacks, and capitalist ventures, this developed AI system called OpenAI threatens to revolutionize the dissemination of convincing fake news stories at an alarming rate.
An AI-written news article is generated as a robotic undertaking, working like a seemingly legitimate news report done by a professional writer. With just a few seconds to complete an article, this means batches of automated articles can get produced in the scale of a big factory – or even more.
Although existing algorithms have long produced basic web articles from minimal data since many years ago, OpenAI takes the process a step further. According to Yahoo, the AI system undergoes a training process, then it is simply fed a sample text providing just a few significant information. From there, the AI-based automated text generator starts composing coherent paragraphs that could have believably been written by a human. It can even adapt to the style and content of the prompts provided, to the point of picking up on the fitting tone for the write-up based on the theme or topic discussed.
In the OpenAI-released sample that features a news report involving unicorns, the technology takes sensationalism to a whole new level. It makes qualitative data sound like an easy mechanical job, inventing a relatively believable fake news write-up resembling human writing – complete with reasonably sounding quotes from credible people and source materials. It is quite clear that its use will eventually yield to misleading and abusive posts allowing turbo-charging fake news through automated publishing, as well as providing more efficient generation of phishing emails, impersonating people and institutions, and harassing individuals more than what existing libelous attacks are capable of. They can also power a presidential campaign story to sway voters and even initiate hate crimes and get away with them, since no human being is directly involved in the process. Indeed, the risk of malicious use of data becomes so high that it can eventually flood media platforms without much effort but with definite results.
Other than the issue on replacing professional writers and journalists for the work they render as the AI system now has the ability to deliver convincing articles at lightning speed at a very low cost, there is a big concern on making fake news stories spread like wildfire in pretty much all possible online domains. The technology can also be weaponized similar to how bots are used in social media for surveillance capitalism. Since these bots are software programs and not real people, the likes of Mark Zuckerberg can legally wash his hands from cases of abuse involving Facebook users’ personal data. In the case of AI software fabricating an endless supply of fake news stories that spread like spam in countless websites, they work like a virus with the potential to confuse people on what’s real and what’s not in a disturbing fashion.
Considering such threats, OpenAI decided to “only releasing publicly a very small portion of the system’s sampling code.” However, it is pretty clear that this AI system will eventually be part of the future. It is inevitable that the time will come that highly specialized algorithms will find their way to the mainstream that governments should be equipped in systematically monitoring and regulating their societal, cultural, and technological impacts.
While this powerful AI technology continues to instigate fear on what the future can bring, combating it doesn’t mean preventing its eventual (inevitable) coming. The legislation and regulation needed to ensure the welfare of the public require the use of a reliable technology of at least the same caliber – perhaps a counter-AI technology that can efficiently and fairly police fake news spread throughout cyberspace.
You have to know the monster and learn how to be one in order to beat it.
Soriano, R. (Graphic Artist). (2019, February 22). The Proliferation of Fake News [graphic art]. Graphic art made using Adobe CC with royalty-free photo from Pixabay. Manila, Philippines: Artist’s graphic creation.
Parkinson, H. (2019, February 15). AI can write just like me. Brace for the robot apocalypse. The Guardian. Retrieved from https://www.theguardian.com/commentisfree/2019/feb/15/ai-write-robot-openai-gpt2-elon-musk
Meer, A. (2019, February 18). Elon Musk’s AI company created a fake news generator it’s too scared to make public. Yahoo! News. Retrieved from https://news.yahoo.com/elon-musk-ai-company-created-195230351.html
Hackett, R. (2019, February 23). Cyber Saturday – Rise of Surveillance Capitalism, China and Iran Go Hacking, Facebook as Digital Gangster. Fortune. Retrieved from http://fortune.com/2019/02/23/surveillance-capitalism-book-facebook-gangster-china-iran-hacking/
Kahn, J. (2019, February 15). The AI That Can Write a Fake News Story from a Handful of Words. Bloomberg. Retrieved from https://www.bloomberg.com/news/articles/2019-02-14/the-ai-that-can-write-a-fake-news-story-from-a-handful-of-words
Bershidsky, L. (2019, February 15). Artificial Intelligence is Getting Good at Fake News. Bloomberg. Retrieved from https://www.bloomberg.com/opinion/articles/2019-02-15/artificial-intelligence-is-getting-good-at-fake-news
Indo-Asian News Service (IANS). (2019, February 15). Elon Musk Funded OpenAI Decides to Hold Back AI Software that does Machine Translation. News18. Retrieved from https://www.news18.com/news/tech/elon-musk-funded-openai-decides-to-hold-back-ai-software-that-does-machine-translation-2037249.html
Radford et al. (2019, February 14). Better Language Models and Their Implications. OpenAI Blog. Retrieved from https://blog.openai.com/better-language-models/
@OpenAI. (2019, February 15). OpenAI. [Twitter post]. Retrieved from https://twitter.com/OpenAI/status/1096092704709070851
#OpenAI #fakenews #automatedfakenews #AI #AIsystem #technology #artificialintelligence