Back in 2016, headlines like these started popping up more and more often:
“Google’s ‘DeepMind’ AI platform can now learn without human input”
Translation: “Google researchers introduce new form of memory-augmented neural net”
“IBM Robot Watson Made A Movie Trailer & Its Success Should Be A Little Scary For Us Humans”
Translation: “IBM Used Watson to Suggest Scary Clips To Be Put Into A Movie Trailer”
“Google’s AI created its own form of encryption”
Translation: “Google Researchers Apply Existing AI Algorithm to Cryptography”
At the time, AI hype was still building towards its current fever pitch and I was already getting increasingly frustrated with this trend and wanted to do something about it. Having taken many classes and done research in AI, it was shocking to me how many such flagrantly wrong headlines got published and the frequency with which they got published. I thought about writing a blog post titled “AI Headlines These Days Are Wildly Inaccurate - So I Corrected Some”, but quickly abandoned the idea since the deluge of misinformation was too massive for one blog post to have any real impact. So, I just gritted my teeth with each successive crazy headline.
Why do I get so annoyed by these headlines? After all, this happens all the time in journalism, it’s nothing new. Here’s why: the societal impact of AI is growing faster than ever, and everyone (including non-AI researchers) will be affected, and so everyone needs to have an accurate understanding of the scope and nature of this impact. This understanding will not come easily, since AI is easy to make up stories about and sensationalize; we are intelligent yet do not understand our own intelligence, and consequently wonder when our creations will surpass us. Accordingly, our culture has built up a century’s worth of fanciful science fiction portrayals of futuristic AI, and present day media now routinely interprets factual research through the lens of science fiction. But the fruits of AI research are now very real, and it is important that peoples’ perceptions of them are based on fact rather than fiction.
Still, I thought I was powerless to do much about it, and so sat by and waited for things to improve. A year passed, nothing changed; actually, things got worse. I started seeing ever more absurd headlines such as “An AI backed by Elon Musk just ‘evolved’ to learn by itself” and “Facebook AI Invents Language That Humans Can’t Understand: System Shut Down Before It Evolves Into Skynet”. And then, that latter headline was so annoying and the story was so pervasive it gave me a sudden idea:
And here we are!
But Are Others Not Doing This Already?
Newsletters like This Week In Wild AI and Import AI do a fantastic job of accurately covering exciting developments in the field. Likewise, sites such as Wired and The Verge do a good job of deflating innacurate stories every once in a while, but they focus on many more things besides AI. Lastly, AI researchers such as Andrej Karpathy and Denny Britz have on occasion written detailed blog posts putting some of the biggest AI events into perspective. All of these are great resources, but it remains the case that no publication today specifically and solely focuses on deflating the increasingly sensationalist and often downright misleading coverage of AI. Something akin to an AI-specific Snopes needs to exist.
And that is why you are reading these words now. From the humble origins of an abandoned blog post and an impromptu idea shared on Facebook, this now Skynet Today, a site solely dedicated to providing accessible and informed coverage of the latest AI hype and panic.
The contributors to this effort will be people like me: not professional writers, but technical people involved in AI who want to contribute to others’ understanding of it. We will seek to promote this understanding through short, easy-to-understand summaries of recent major media stories such as those already on this site, as well as longer editorials aimed at discussing overall trends and ideas. None of the contributors will profit monetarily from these articles, and all will do so in their free time motivated by the simple belief that someone qualified needs to explain what’s really going on to a general audience.
So, if you agree with this site’s mission and are such a qualified person, get in touch! We are currently looking for additional contributing writers. And if you are not such a technical person, but a person interested in this content, get in touch! As we are just starting out, feedback and ideas are invaluable.
We believe understanding of AI and its ramifications should be more accessible, that some of those with the technical knowledge to seperate fact from fiction should help make it accessible, and that those unsure of its ramifications should seek to become more informed. Plus, we won’t be shy about poking fun at the sheer absurdity of the stories we will criticize, so we can all have some fun in the process :) .