CS123, Intro to AI
Topics | |
---|---|
Overview of AI | Neural networks and deep learning |
AI Problem Solving Revisited Machine Learning—Part 1 Applications of AI | Generative AI + Prompt engineering |
Machine Learning—Part 2 | Custom chatbot creation |
History of AI + Midterm | Social and ethical issues Final |
IntroductionWhat's Happening this WeekOverviewPotential benefits of AIPotential Dangers of AIWhat Can Be Done?By IndividualsBy OrganizationsOrganizations that create AIOrganizations that use AIBy GovernmentsReferenceArticles and TutorialsWeb Sites
Due Friday, 8/16 (the last day of the term)
Forum post with questions, answers and comments for online students (Part of your participation grade)
Exercises based on Elements of AI Ch. 6, Implications
Final quiz
In the classroom, Thursday 8/15 at 10:00
In the testing center, Wednesday 8/14 and Thursday 8/15 from 9:00am to 5:00pm
Increased work productivity
New business and job opportunities
New discoveries in science and medical fields
Learning resource in education (as a tutor)
Improved cybersecurity, law enforcement
Empowerment for less privileged people (can generate content without high cost)
Companionship for elderly or lonely people (a benefit touted by some, but questionable)
AI used for defense and law enforcement
Lost jobs
Invasion of privacy (by governments and corporations)
Violation of copyrights, no attribution or remuneration to authors or artists
Hindering education (students short-circuit their learning)
Diminished sense of what it means to be human (machines that appear to be persons could lead to persons being regarded as machines)
Environmental damage due to use of power and water by data centers
Bias in training and inequitable access
Spreading false information, deep fakes
Used for phishing and cyberattacks
Military and police abuse
Humans loosing control of AI
We can take steps now to steer the course AI takes to direct it toward good instead of the other thing.
We vote with our fingers. By using an AI platform we promote it. Choose the ones that are more ethically trained, environmentally sustainable and responsibly managed.
None of the big four AI systems are profitable yet, as they find ways to monetize their AI, users will have leverage to influence by choosing what they use.
Use AI in responsible ways.
Don't become over-reliant on it.
Don't use it without attribution (don't pass of AI creations as original human work).
Don't use it to create inappropriate content.
Protect private data (be careful what you put into prompts or RAG data stores)
There are four companies that currently dominate the generative AI market with their LLMs:
ChatGPT by OpenAI which is partnered with both Microsoft and Apple
Gemini by Google
Llama by Meta, Facebook, Instagram and What's App's parent company
Claude by Anthropic which is partnered with Amazon
None of the big four LLMs have been certified as being fairly trained.
All of these companies claim to be implementing safeguards to keep AI "safe" and to deploy it "responsibly", but there have been allegations that they all have been too hasty in rolling out their systems.
Establish policies so that employees have guidance on safe and beneficial use of AI. Here's an example of areas being addressed by the LCC AiTF:
Intentional incorporation of AI in the classroom. Use AI as a teaching aid and teach students how to use it in ways that don't "short-circuit" their learning.
Safeguarding academic integrity. Prevent cheating and promote transparency by identifying AI created work.
Keeping data secure in light of new technologies.
Protect private data from being used for training, RAG, or incorporation into prompts
Addressing pedagogy and student assessment.
Look for positive uses in these areas.
Exploring new academic programming and fields of study that include AI technologies.
Add new courses, certificates, or degrees.
Keeping abreast of developments of AI and responses of other community colleges and higher education institutions, including accrediting bodies, or in other words, develop an expertise in the area of AI.
Ensuring and promoting equity, diversity, and inclusion in this new technological landscape and specifically utilizing Lane’s Equity Lens in all deliberations.
Establish laws to regulate AI. It's not sufficient to have AI companies regulate themselves.
Use AI in responsible ways. Limit its use by the military and police
Here’s Proof You Can Train an AI Model Without Slurping Copyrighted Content—Kate Knibbs, Wired, 2024.
Regulation of Artificial Intelligence—Wikipedia
Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence—White House, 2023.
LCC AI Task Force—Committee working toward policy recommendations for the college
AI for Good—Initiative of the ITU, a UN organization
Fairly Trained—Certifying ethically trained LLMs
Intro to AI lecture notes by Brian Bird, written in , are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Note: GPT-4 and GPT-4o were used to draft parts of these notes.