Forum

In the interview with Ben Shneiderman, one of the prognostications was a warning that: "Machine learning, neural networks have great opportunities, but the mindless acceptance of their potency is extremely dangerous." I think this warning has very much come true and will continue to be very important. For example, there have been a number of crashes of self driving cars resulting in injuries and fatalities. On a smaller scale, we are starting to all become more accepting and reliant on AI, leading to the spread of misinformation. I have seen a number of AI deepfakes that are completely fake pictures of famous people in certain scenarios, but with the help of AI these deepfakes have been easier to create and much more convincing. Even scarier I've seen some videos that become more and more convincing as well. I think Shneiderman's warning will be really important moving forward as it is super important to think critically about our AI usage and ensure we're creating enough regulations to ensure we progress forward with safety.

During this interview, I was fascinated by the difference in machine learning trends and statistical trends, especially the comparison with AWS. However, the discussion around machines as tools rather than partners was interesting. They discussed how machines would be able to mimic many actions that humans take and do what humans do, including composing music. They claim that AI composing music is still very far away, however, AI agents including Suno and Mureka already compose music based on user preferences with AI tools. AI is currently used as a tool extension, with ChatGPT and DeepSeek being used for many different tasks to make human lives easier. The question that was posed, how will we be able to tell the difference between a smart human and an AI, is important as AI continues to progress at a rapid pace. AI was being used as a tool in 2018, but one can make a case about over reliance on these tools by people to make their lives much easier and replace many jobs. ChatGPT is able to write similar code as an entry level developer and can solve the hardest leetcode puzzles, showing how AI is starting to not only be used as a tool, but actually as a replacement. This shows that he was correct about a few things, including how AI will be able to mimic human actions and will be used as a tool, but counterarguments can be raised regarding the speed that AI has developed, where AI can replace many entry level jobs and even compose music, a task that was considered to be very far away only a few years ago.

We’ve always built tools to make life easier as Sheridan notes. For most of history, that meant offloading physical work or boring, repetitive tasks like the bulldozer example. These tools gave us extra muscle or saved time on tedious chores, but we still had to do the thinking part. AI is different because it takes on the thinking itself—the cognitive tasks that used to require a person . You could use a past tool like spell-check to fix typos, but an AI can, and already does, write a whole paragraph for you. That’s a fundamental shift, and it’s why AI feels less like a simple helper and more like a new kind of partner (or sometimes a competitor) in our work.