uses of AI
Generally
Section titled “Generally”- is this something that actually requires a more efficient solution than the existing methods?
- eg improving healthcare likely a good application
- but true creative works don’t rely on efficiency
- many uses that are varying viewed are just showing another process should be automated - eg ‘if jobhunters and employers are both using AI what’s the point in the process’ - yes, correct, application processes are often incredibly redundant, skill matching could actually often be automated and send you straight to ‘culture fit’/personality power-play stages of interviewing that are the subjective measures, skipping the unnecessary ‘quantitative’ stage.
-
likely good uses
-
(with hard requirement of proper regulation and large research backing and human involvement)
- detecting illness
- some simple triage
-
very risky, too easy to hijack and difficult to argue for, likely not worth it
- any human image recognition
-
pretty useful applications that could likely use more specialised models/training, could pause until more ethical power consumption
- all-in-one language tool
- synonyms, antonyms, tip-of-your-tongue
- general etymology
- cross-language tutoring
- relevant examples, tuneable
- summaries
- tuneable complexity and approach
- request more concise answers
- asking for clarification of particular parts
- tailored language for particular audiences
- cultural relevancy and context
- explaining metaphors
- explaining idioms
- explaining jokes
- enhanced databases
- better search within a very specific smaller kind of data
- all-in-one language tool
-
neutral (may pose ethical issues, maybe not possible, maybe a distracting research focus, maybe not worth it until more ecological considerations in the tech, maybe there are other adequate solutions already, or I’m generally not sure where to place this)
- word processing, formatting
- general search engine
- for more conversational answers on simple info just integrate small conversion models into existing search engines
- understanding nuances
- explaining possible meanings or intentions of certain statements, given knowledge or lack of knowledge or the general experience or values of a speaker
- large amount of context needed
- socratic dialogue to work through your own thoughts
- non-judgement
- allowing you to rationalise quicker in emotional circumstances
- for when you lack immediate contacts with the necessary temperament
- adjacent to therapy-bots
- case studies
- wider reach than other collated examples
- what people ‘tend’ to do
- external factors
- very general legal advice
- some quick options of actionable steps in commonly faced issues
- learning from others’ mistakes
-
we don’t yet have another adequate solution but AI is still probably not it
- search and discovery of information, projects, people, works, etc
- interdisciplinarity
- vague questions that span historical, scientific, cultural knowledge etc
- eg recently I’ve searched - how long have dogs been pets - how did they first make bread - why is your brain not in your ribcage or your heart in your skull
-
absolutely not
- any creative work
- very rough ‘sketches’ of concepts can be useful to an artist’s workflow but the associated AI-creep into creative industries often isn’t worth the tiny-time-saver that is generating concept imagery - not sure about this
- for all small applications that are said to ‘democratise’ art, eg a quick render of your D&D character concept
- artists on UBI throwing out free customised character concept works of their own volition is so much preferable to stolen work in AI image gen creating ugly untuneable creatorless avatars that don’t really look how intended anyway
- art is inherently ‘democratised’, capitalism denies this from artists who need to survive and non-artists without the skill or time to learn or the funds to support the artists to do it for them. X