Do robots dream of electrical sheep?

Digital health

Digital health is a multidisciplinary field aiming to improve healthcare delivery with IT in terms of quality, efficiency, equity and/or creating new opportunities. It also involves the management and use of patient health care information. Applications include:

• AI in medicine (e.g., expert systems, decision support systems).

• Electronic medical record (EMR) systems.

• Information systems such as implementing large-scale, organizational change, data coding standards and interoperability.

• Use of IT by health consumers.

This topic can be seen to also include bioinformatics — computing about genes, proteins etc.

Case study: CPOE

Drug interaction checking

Alert fatigue

Define

Doctors will quickly tune out (and turn off) if an interaction checker fires false positives (FPs).

AI and workforce

A direction of thinking in AI replacing jobs is: do you actually do anything? A lot of 'boilerplate', background, definitions and other padding can be done by LLMs like ChatGPT — an example being generative AIs' capability to summarize.

Still, there is tremendous risk in letting an AI do your writing for you (if what you write matters at all). Take the use for legal/paralegal work as a use case. What if it makes up the cases that underpin your legal brief?

Applying this topic in the art/illustration workforce, there used to be a lot of demand in creating images for advertising and other illustrations, cover art, etc. Generative AI images are slashing into this industry, leading to companies needing only 'touch up' workers.

Bringing up another example, work of journalists can be accelerated by generative AI in many ways:

• Writing the final product: Main text, summarizing sources, headlines, etc.

• Research and fact checking: Gathering information in prep to write an article and verifying claims of others.

They could even have fully automated pipelines that produce news feeds without human intervention on individual posts, potentially leading to better journalism by easier/faster execution of tasks and/or worse by distortions and errors. Most likely, workforce reductions over time since the generative-aided workers will be faster.

Use case: Deepfakes

One application of generative AI is replicating voice actors,resurrecting or de-aging individuals. This process is sometimes done with consent — as is the supposed case with Darth Vader voice actor Peter Cushing. However, this can also be done based on obvious potential to reduce need for actors or reduce pay (e.g. for stand-in actor), although more negative potential to be done against wishes of actor or their estate as well, depending on legal context.

A currently more framework-universally, 'bad' usage of vocal generative AI is Deepfake creation — a video of a person in which their face or body has been digitally altered so that they appear to be someone else. It is an extension of false-attribution and image editing that has always been possible, but the ease of achieving reasonable quality fakes has improved dramatically with the rise of generative AI quality, decreasing the skill ceiling for creating them.

Use case: high demand in healthcare

AI in managerial work