Reflection 5
This week was a very exciting and fun week because Helen Armstrong visited us! It literally felt like a celebrity was coming to our class since we have talked extensively about her work throughout the semester. In our Monday class session we listened to Helen’s lecture on AI assistants in the past, showing that this isn’t necessarily a new idea. I really enjoyed when we looked at clips from movies like Space Odyssey, proving that humans have been thinking about these ideas surrounding AI agents for a long time. What noticeably stood out to me was that it gave the AI so much control, where it actually stops the human from doing something and it is too smart to be tricked. I think that definitely raises some questions and points of interest that should be considered while developing these agents.
Helen’s lecture on Tuesday night.
I wasn’t knowledgeable on the difference between AI agents and assistants until class, where we learned that an AI assistant could be considered an agent, as long as it is able to pursue goals without specific instructions, using tools it has and able to understand natural language so it can act automatically. Within this category we then covered the two different types: a tool-based agent and a simulation agent. It’s kind of weird to me to think about these simulation agents since I feel more comfortable with the tool-based agents. Human-like behavior is not easy to copy, so I’m interested to see where these will go, especially the freaky avatars that represent you, as if you were on a meeting or something. Helen mentioned that this would allow you to get more work done, your avatar would take your place and you could tell it what to say, but I’m confused how that would work since conversation doesn’t work like that. Often times, I enter a meeting where I know what I would like to say, but conversation swerves frequently, jokes are made, we get off topic, etc. There’s no way to predict everything that will happen, so wouldn’t you need to be watching your avatar anyways so that you could tell it how to respond? And then what is even the point?
On that note- I’m not a huge fan of innovations that just push hustle culture even further… Yes now you can be on a meeting while also managing a staff and completing other tasks and responding to emails, etc. You see where I’m going with this. That’s not a way to work or live in my opinion. Meetings are a way to stay connected to your team and coworkers so I’m not a huge fan of all of us sending these digital representations of ourselves to them. But maybe I just don’t understand quite how it would work yet. I enjoyed our discussion about how the use of agents could go wrong because it makes me feel better that people are so perceptive to the effects AI will have. This will make it easier to implement it safely into society.
Ring mock up.
Helen’s lecture on Tuesday night was a really great synopsis of everything we have read and talked about in class. I honestly left the talk and dinner feeling much more empowered to play with AI and try to jump on this train instead of resisting it. I always love hearing about how designers are needed because of the specific skill sets we have; our training to seek a desired future or right thing to experience. I used to always worry about going into design when I was younger because I didn’t feel that my job would be fulfilling since I wasn’t making an impact in the way that someone in healthcare does. I understand now what an immature way of thinking that was because designers are the one that ensure humans are in the center of the designs we interact with. It’s cool to hear how our work can change so much! A point she mentioned in her talk that really resonated with me what that humans tend to overtrust AI. I have known this to be true for myself in the past… I’ll ask Chat a question and if it is wrong I feel a sense of anger… how could this language model possibly get this question wrong?!?! And then I will feel like I learned my lesson and will not turn to Chat again for a similar issue. Helen discussed this exactly, that humans will trust AI right up until they question a predicition and then will struggle to build back up trust. How can we create designs that encourage skepticism but also provide recourse when there is a disagreement?
App mockup.
One last point from her talk that I think will always stick with me is her mentioning the creation of computers and the general distrust and fear surrounding that creation. Now, primarily everything I do for my career is on my computer! It really helped me change my mindset with AI and I really want to try her suggestion of just getting in programs and playing with it so that I can be ahead of the learning curve because AI will soon become a bigger part of our everyday life. We talked about this at dinner but it is so hard for me to fathom what my job will look like in the next 5-10 years because we will be designing for things that don’t exist yet. Although it scares me, how exciting!
App mockup.
In regards to our workshop in class, now moving on to Wednesday, Helen challenged us to create the worst AI agent possible based on a script we were presented with. My group was assigned a script regarding productivity, as the user wanted something to ensure they completed their tasks without the situation presenting itself as important without the reminder feeling evil. I really loved getting to ideate with my group over a bunch of different ideas and get creative regarding the craziest, most ridiculous idea we could think of. I really struggled towards the end of sketching out 8 different ideas, but it did really push me out of my comfort zone. I had fun getting to design the hypothetical ring that would light up and buzz if you didn’t complete your task, as well as getting to incorporate some UI design practice. I really liked getting to verbally discuss how the AI would be problematic because as we talked more and more ideas kept coming to mind. I think the ring actually would be a really useful tool for someone like me who relies on being held accountable and enjoys a challenge. However, there definitely would have to be some tweaking done to ensure it was actually a useful tool and not a complete annoyance, like including an override option or silence mode! Seeing everyone’s ideas and thought processes was a really nice way to exercise critical thinking when it comes to AI usage and it made me realize that I know more then I think I do, because whenever someone presented an idea my mind would immediately start analyzing what was good, what needed work, and what implications came with the product. I do really think my mind is being trained to think in a completely different way from when I began the semester, which really excites me. We wrapped up the workshop by going through a last discussion, asking questions, and picking Helen’s brain just a little bit more. Overall, this week and the workshop left me really inspired to play with AI and just imagine what it can do!
