Mike Riley - AI Toolbox and Guest Lecture at AI
LIFE WITH ARTIFICIAL INTELLIGENCE AND AI
Undermining the Public's Trust
AI Imagery such as deepfakes, I firmly believe both undermine the public's trust and also take credibility away from the public's trust. Based on many of the journalistic integrity principles I have learned from throughout my time at Elon. Though tools such as data wrapper and diagram.io will be extremely beneficial for people in newsrooms who are data, enterprise, and investigative centric, I feel, as the Associated Press says that 20% (in my opinion preferably more) should be done by humans.
Best Practices Going Forward for Regulating Artificial Intelligence
With over approximately 2,000 new AI tools and platforms come out on a monthly basis, to me it seems that there is no possible way that 1 individual, government or regulatory organization could possibly be able to keep up with and keep track of what is good or bad from a utilitarianism perspective.
_____________________________________________________________ CLASS NOTES
AI JOURNALISM TOOLBOX
HOW AP USES AI
80: 80% of process can be automated
20: 20% requires hand done fact checking
AI: A Journalists Approach
Traget repetitive tasks
AI Tools the bad:
easier to create misinformation and disinformation
Where is my data going?
Ai=I Transcription tools and concerns
Otter
Terms of Service: allows the company to take and use the data
Otter, benefits and drawback of transcribing and logging by hand.
(HOW HONEST ARE THESE COMPANIES SUCH AS OTTER WITH THEIR USERS ABOUT PERSONAL AND PRIVATE INFORMATION
——————
Generated AI: Biases
ask to write a love story and it will write a hetero based love story
Asked Stable diffusion to create a picture of an inmate output: is a
____________________
Where copyrights stand
Any works using AI may be copyrighted given that the content, piece includes human
March 15, 2023: Copyright law decision
Copyright Laws of Concern
Fair Use
allowance is made for fair use for purposes such as criticism comment, news reporting, etc
MidJourney V6 has cleaned up “tells” on fake images
Discolosing AI Use to Audience
when of you tell readers that you used Ai to produce something?
Do you feel readers and audience that you spell checked something?
Ethics Guidelines from SPJ NPPA, PRSA, TrustingNews.com
Using the term art director when published to the public in disclosing AI USE
____________________________________________________________
Best Practices to Achieve Results and Outputs for Journalistic Settings
How would you like the AI to respond/what you’d want
Tell the software how to treat me or the data enterer
Using AI to find databases, with reliable online links?
assign it a role tell the software to play the role for me
_______
Databases
Open.Ai
diagram (creates visually appealing graph
Hemingway App (cleans up grammar for writing and journalism: software to give you editing scores and to help you write?
AI and Plagarism detection
GPTZero
summarize.tech : will sumarize longer-form youtube videos
video.yo
open.ai.com
openai.com
video.ai
Deep dream generator
_______________________
Struggle with verification of official data and statistics, specifically user when user error in inputs.

















