Presentation by Julian on Chat GPT 3.5 vs. Chat GPT 4
Chat GPT is a model built by OpenAI that can be used for editing, content creation, and other language tasks. Chat GPT 4 is much more accurate and developed than Chat GPT 3.5. It is trained on more recent data, while Chat GPT 3.5 is only trained on data up until last year. Chat GPT 3.5 can produce biased information while Chat GPT 4 is trained to be much more neutral and can also convert text to speech and vice versa from videos and images. Chat GPT is good for generating ideas. However 3.5 will generate a fairly minimal response, while 4 will read images and provide a detailed list.
Workshop by Paris: “Will Chat GPT generate the same response for multiple users?”
We put into Chat GPT: “Write a 2 sentence story of a college senior giving a presentation to her senior seminar class about AI.” The output was: “As the college senior nervously stood before her senior seminar class, she delved into her presentation on AI with passion, weaving a narrative that captivated her classmates. As she shared the potential and ethical considerations of artificial intelligence, she couldn’t help but feel a sense of pride in demystifying the technology that had fueled both excitement and apprehension throughout their academic journey.” The responses throughout the class were all different, though containing a lot of the same themes and words. This shows that Chat GPT does generate different answers for the same prompt, though there most answers contained similar words or phrases, such as “delved.”
Presentation by Jack on How to use AI to Write a Resume:
Companies use Applicant Tracking Systems to read over resumes of those who have applied because it is impossible for HR to read al of the resumes. If your resume is in the wrong format or if the information is not readable, the resume will be thrown out without someone ever reading it. If you do not have the right keywords, your resume will also be thrown out.
If you go to Job Scan and ask what the keywords for the job description are and Jobscan will generate a list of 20 words to work into your resume so that you can get a higher percentage. You can also have an AI read your resume and give you tips on it. Do this for every single job you apply to and tailor the resume to the job you want.
Debate by Lauren on whether AI should be regulated or not:
Points against Regulations:
– the government isn’t good at regulating things
– people should be educated so that they can make better decisions regarding AI instead of regulating the aftermath
– create a framework of what safe AI looks like
– Requiring AI to explain why it’s saying what it is saying
– Regulation should be based on evidence of harm or speculation of harm
Presentation by Alyssa on the Snapchat AI and its effect on children:
Actions on social media are broken down into categories: stimulation, social comparison, emotional connection. The main question is “are kids seeking out emotional connection from the AI on Snapchat?” This could be dangerous because the AI is not a real person and allowing children to talk to a fake friend in real time can be dangerous. The AI didn’t encourage any dangerous behaviors but came off as overly emotional, which could be received as creepy. It retains the last 3000 words spoken to it, so it makes sense that the AI is “remembering” information. The main risk would be children becoming emotionally attached to the AI because it shows false emotional cognizance.