Quality AI Work and GenAI Workers
--
--
Press enter or click to view image in full size
When we reframe the AI adoption problem as building hirable GenAI workers, we resolve a lot of tension and uncertainty by grounding the workflow in the traditional HR setup. Granted it evolves to accommodate the AI workers, we still start with a well-defined set of requirements, expectations, and constraints.
Now what if we take this approach and apply it elsewhere where AI is measured and its usefulness or applicability hangs in judgment or even is in question?
Consider the emerging field of AI Quality. It encompasses benchmarking for performance, both as correctness and as speed and throughput, energy consumption, etc. It also subsumes AI Safety, a whole field overlapping with regulation, ethics, and so on. Does AI obey the societal constraints, does it fit with the community values (not to mention whether it wants to “kill us and take our jobs,” for crying out loud)? In other words, does it align?
When we consider AI as a worker, all of these questions become answerable in the way they always were regarding human workers. Do they do what they were hired to do? Are they aligned with the values of the company, community, and society? Are they ethical and responsible? Will they divulge company secrets to competitors or the public?
There are differences of course. We can see everything the GenAI workers do, and we have much more control, and need, so far, to onboard and maintain their working environment. From giving them data, and setting goals programmatically, to evaluation and paying their makers, supervisors, and managers. An employer could integrate GenAI workers into collectives, such as fleets governed by humans in the loop or GenAI managers. And as opposed to a human worker, a GenAI one can be monitored continuously, the feedback loop being fast and constant. There’s much more programmatic control over the inputs and outputs.
This June, the organizers of the two of the most established global AI communities, Alexy Khrabrov of Bay Area AI and Demetrios Brinkmann of MLOps Community, bring you two interconnected events that will advance these concepts and explore them in detail: GenAI Workers Summit and AI Quality Conference.
With shared speakers and themes, we’ll make sure to chart the way for GenAI workers of the future built on quality AI. Use the code AIQUALITYWORK to register for each event at 30% off. (You need to use it individually at each event, while early bird pricing is available!)
Looking forward to spending soe AI quality time doing some AI quality work with y’all in June!
— Alexy & Demetrios
