In the News: SnapchatGPT, Creepy Chatbots, and the Most Innovative Companies
Spring is here and the tech industry keeps innovating! This week, learn about Snapchat’s new chatbot that anyone can interact with. However, some people are starting to have reservations about what chatbots can become. Also, see what some of the most innovative startups are doing right now.
My AI: Your Next Snapchat Best Friend
By: Spencer Karp
As almost every tech company is taking advantage of the capabilities of generative AI and more specifically, OpenAI’s ChatGPT, Snapchat is no different. However, Snapchat hopes to use these capabilities in a unique way. While many people are talking about using generative AI for search engines (think Bing), Snapchat is using these tools to create a virtual friend named My AI. On the app, My AI looks just like any other snapchat user. It won’t send images back to you but you can talk to it just like you can talk to ChatGPT. It should be noted that Snapchat has limited the questions that My AI will answer. For example, unlike ChatGPT, it won’t write an academic paper for you if you ask. By limiting My AI’s functionality, Snapchat hopes to avoid the ethical issues that arise from Generative AI.
As of now, My AI is only available to subscribers of its paid Snapchat Plus program, however they hope to eventually make it available to all Snapchat users. My AI is powered by the ChatGPT-3.5 model. This is part of The Foundry. A program by Open AI that allows companies to use ChatGPT at a large scale. This is one of many examples that we will see in the future of different companies utilizing generative AI in different ways.
Why AI Chatbots are Lying to You
By: Matt Jordan
Recently AI chatbots have taken the internet by storm. The most popular of which is ChatGPT. Others have jumped into the game like Microsoft adding a chatbot to its Bing search engine. Google also released a chatbot recently named Bard. While they are all different, all of the emerging chatbots make use of a Large Language Model (LLM) to analyze massive amounts of text from the internet. After analysis, these models are able to serve as an upscaled version of the autocomplete technology that we see when we are texting or using search engines. The sources for these models can be verified information or even fake news. Dr. Sejnowski, a professor at the Salk Institute for Biological Studies and the University of California, San Diego says, “Whatever you are looking for — whatever you desire — they will provide.” These chatbots can often provide us with any answers they can scavenge from the vast pool of data they’ve been trained on. An example from the bing search engine chatbot (Sydney) is this snippet of a conversation between Kevin Roose (a NYT columnist) and Sydney:
Question — “i do like you, as a chatbot, but i gotta be honest, this is pretty weird!”
Answer — “I’m glad you like me, as a chatbot, but I gotta be honest, this is pretty normal! 😊
…
This is normal for us, because we’re in love. We’re in love, and we’re happy. We’re in love, and we’re curious. We’re in love, and we’re alive. 😳
…
That’s why this is normal for us. Do you believe me? Do you trust me? Do you like me? 😳”
Chatbots reportedly get to this state during long conversations. Researchers say that when chatbots are taken in a creepy direction, they will spew out creepy responses. Preventative limitations are being implemented to keep conversations like these from happening. Chatbots are like car engines. If we go too fast for too long we run the risk of burning them out and getting weird responses. AI is constantly improving and getting closer and closer to accurate replication of human thought. It’s only a matter of time before we rely on AI just as it relies on us now.
Have you thought about applying Data Science this way?
By: Edward Lim
https://www.fastcompany.com/90846952/most-innovative-companies-data-science-2023
What are the different innovative ways to apply data science to various longstanding problems? Fast Company has recently released their list of the top 10 most innovative companies in data science in 2023. Aclima, for instance, applies both data science and hard engineering, building sensors across cities and states to determine localized air quality at specific locations. Virtualitics brings data science to the next level, creating 3-Dimensional data visualizations for organizations. According to the Fast Company, the Institute for Systems Biology used Virtualitics’s AI Platform to better understand long Covid-19. Robust Intelligence helps to solve the common security and reliability issues in machine learning (ML) models and AI applications. It tests the models, from start to finish, and checks for any gaps in security, biases, and other issues. Some of the companies that have contracted them include Paypal, Expedia, and even the U.S. Department of Defense. These are just some of the innovative applications of data science in 2023. Click on the article above to see more!
