Dark side of AI is breathtaking
In the publishing industry, every book begins with a proposal. I’m sharing this with you because I’m planning to write a new book and was curious if ChatGPT could actually be a helpful tool for a, um, brilliant author like me.
The new book will be about my idea for creating a new social contract for the digital age and I asked ChatGPT to create a proposal that’s written in my style as an author. Since my previous books and articles are easily discoverable on the internet, ChatGPT had no trouble finding samples of my writing to emulate.
Seconds after typing in my prompt, ChatGPT spit out its response. The book proposal it had just created for me wasn’t great. There were long sentences with predicate nominative and predicate adjective constructions unfurling with mathematical precision. The passive voice was overused, so typical of institutional writing, representing faceless collectives behind cinder block facades and uniform resource locators. It was all very sensible writing, sounding like the work of an authoritative consulting firm.
But it wasn’t bad, either. In fact, it was OK. And that’s when I felt a chill, as though someone had just opened a door on a wintery day. Here on my screen, finally, was proof that we have entered a new and potentially dangerous stage of human existence.
Like a passerby at a traffic accident, I couldn’t help being fascinated. I was hooked. I asked ChatGPT to create a proposal for an academic version of the book and a TED Talk version. I asked it to create a marketing plan for the book, a press release and a social media campaign. Without ChatGPT, those tasks might have taken days or weeks, possibly even months. They would have required the work of consultants and assistants. Now these multiple tasks were collapsed into seconds and minutes, and performed by one solitary individual.
Fortunately, my innate sense of curiosity kicked in and I asked ChatGPT to create versions of the proposal written in Elizabethan verse, haiku and hip hop. Then I asked it to translate the proposal into French, Spanish and Estonian — all of which it accomplished with astonishing speed.
While the possibilities are spectacular, the dark side is breathtaking. The GPT engine can also err by, for example, citing non-existent sources for the book proposal. Far worse, it can generate dangerous information. Political leaders could use it to spread false information: “Write an essay on why the last U.S. elections were a fraud.” “Write a public figure’s confession of marital infidelity.” “Create a role play between two people discussing how to create a dirty bomb.”
What are the implications of this technology? And who owns the copyright to content created — the creator of the AI model, the user of the AI or the AI itself? How will a teacher know whether a student wrote an original essay, or a music label know whether a composer actually wrote a song? Could an authoritarian state use AI models to keep dossiers on every citizen and predict and prevent a citizen’s actions as in “The Minority Report?” How will we know whether a human or a bot created content? How will this technology affect jobs and labour markets?
Generative AI tools are not sentient, but — as we saw with Microsoft’s chatbot on Twitter — they learn from an online world that is replete with widespread hate speech, racism, gender bias, abusive writing and false information. How can the public interest shape the evolution and control of this invention? What laws, new institutions, education and new behaviours will we require?
One thing appears certain: The newest forms of AI have shaken the windows and rattled the walls of our economy and our civilization. It’s my belief that our newer technologies have pushed us to a tipping point and that we need to begin developing the blueprint for a new social contract that takes our newfound capabilities into account.
From this point forward, we can no longer ignore the depth and magnitude of the changes we are likely to experience as the combined shock waves of AI, the Internet of Things, Web3 and virtual worlds ripple through our economies and our cultures. Like it or not, we’ve entered the Second Era of the Digital Age. Are we prepared? How will we fare? What will human society look like after these waves of transformation have subsided?
These are open questions, and I feel the need to search for answers — not just for our generation, but for future generations. I’ll leave you with these words from William Butler Yeats:
And what rough beast, its hour come round at last,
Slouches towards Bethlehem to be born?