Everyone is sick of talking about AI.
Everyone is breathlessly excited about AI.
Everyone is terrified of AI.
The power of generative AI and Large Language Models is inarguable. From a technology parlor trick in 2020, we now have algorithms that extract the underlying meaning from language itself, and can be applied to complex knowledge work. Tens of thousands of dollars of legal work are done for a few dollars in tokens; a week’s research completed in a dozen minutes.
These technologies represent a new kind of computing. A nondeterministic computing. Where unexpected answers were a bug in traditional computing, they’re a feature of AIs. And therein lies the paradox: We can’t know for certain if they’re right.
Early results suggest that AI undermines curiosity, job satisfaction, and the very desire to learn and question. Are humans relegated to the role of overseer, signing off on the work of an AI, grounding it in liability laws?
At the same time, perfection is the enemy of good enough. We hold self-driving cars to an impossibly high standard of zero accidents, when presumably “better than human drivers” would be enough, precisely because we haven’t properly thought through the immense ethical issues surrounding nondeterministic answers.
As we programmed FWD50 for 2025, we wanted to go beyond traditional AI conversations. AI, after all, begins with data. And it ends with automation. So we’ll be looking at the “AI supply chain”:
-
Why data matters, and who controls it.
-
Where it makes sense to deploy imperfect things, and where it’s smarter to wait.
-
What automation looks like, what jobs it changes, and what new tasks it creates
Join us on Access FWD50 for a day of practical, no-hype AI conversations you can apply to your work in every facet of the public service.