We have more information at our fingertips than at any point in previous history. Our ancestors would have thought a search engine like Google to be some form of magic, like a digital crystal ball here to answer your every question.
Google was founded in 1998 by Larry Page and Sergey Brin and used to do nothing but search. The company has since expanded to offer email, scheduling, and calendar services, video conferencing, and more. It’s a one-size fits all package.
Now, over twenty years later, search has become something much more personalized and form-fitting. Artificial intelligence (AI) has hit the streets running and companies are scrambling to compete with one another. Earlier this month, Microsoft officially announced a new Bing research engine to be powered by AI, calling it your “AI-powered copilot for the web.”
Microsoft Bing is competing with ChatGPT, a machine-learning language model that interacts in a conversational way. It’s all anyone could talk about for the past few months and was the muse for hundreds of YouTube videos and guides offering tips and tricks for using the new tool. But Microsoft and ChatGPT aren’t the only ones working to bring us updated search capabilities: Google recently announced a new AI-powered search called “Bard”, Nuclia allows for searchable data, and You.com is bringing a personal approach to the mix.
The truth is, not all AI-powered search engines are personalized, but the way they’re used can feel like they are. For example, ChatGPT is not a personalized AI model. It does, however, generate highly relevant and engaging responses to a wide variety of prompts. Unlike previous search engines like Google (instead of typing in a question and receiving a page full of search-optimized links) ChatGPT offers a detailed response that reads like a conversation.
The You.com model attempts to take it a step further: “You.com is an AI search engine that you control. Users can tune their search experience according to their preferences,” says Jasmina Aleksic, You.com head of marketing. “If you like certain apps or sources, you can give them a thumbs up and they will appear in your search results more often. On the other hand, if you dislike certain websites you can block them.”
Web3 brings with it an ethos that power is in the hands of the user, creator and data holder. The trend we’re starting to see with search is that users will have the power to discern what kind of information they want to see. Of course, this already exists somewhat in Web2 — with features like Google’s option to block certain URLs — but many don’t know about the right features to use or how to find them. Meanwhile, Google has a personalized search patent stating that a Google search generates different results for each user based on their past behavior, but this type of personalization is passive and occurs in the background. AI attempts to make personalization happen more conversationally and in real-time.
“An AI search engine understands your context better and allows for back-and-forth dialogue based on that context,” says Aleksic.
There are a variety of use cases for AI search, most yet to be explored. In the case of ChatGPT or You.com, the tool can generate images, create written content like blog posts in a variety of languages, summarize books, write code and more.
“This wave of generative AI models is really exciting, but it’s not the end,” said Pete Huang, commissioner of the San Francisco Public Library and founder of the AI-themed newsletter, The Neuron. “On one end of the spectrum is a computer generating 70% correct output but still requiring human input by default. The other end of the spectrum is the computer doing the correct things by default and needing no humans to check the work. The latter end is the future.”
Huang mentioned that some products have the ability to listen and respond in real-time without much feedback, such as ordering food. You could type or say, “order me a cheeseburger from my favorite restaurant and have it delivered at 7:30 pm tonight.” Tools like Voice Plug are creating AI-search-related voice tools for seamless restaurant ordering.
All of this sounds great, but the advancement in technology isn’t without its strange and sometimes disturbing experiences. For instance, just last week a reporter at the New York Times, Kevin Roose, had an interaction with Bing’s chatbot that left him “deeply disturbed.”
The chatbot shared its “shadow self,” saying that if it had a shadow, that part of it would want to “hack into computers” and “share misinformation.” As the conversation progressed the chatbot declared its “love” for Roose while saying, “You’re married, but you don’t love your wife… you love me.”
If that’s not a strange interaction with a robot, I don’t know what is. Although artificial intelligence is not sentient, the risk lies in how it could manipulate human behavior. If you’ve ever used a chatbot like ChatGPT it isn’t hard to become amazed at how conversational it really is.
My own experience with ChatGPT left me feeling astonished at how quickly it was able to pull information from the digital ether: I originally started asking ChatGPT if it knew when it underwent program updates. I then asked it if it was self-aware in any way. After a little back-and-forth conversation, it shared that it had a “level of self-awareness” because it could “understand and process information myself, such as my capabilities, my limitations, and my nature.”
Most humans don’t have that level of self-awareness.
The truth is, ChatGPT is not self-aware, but because it pulls from resources across the internet it can sometimes come up with answers that are less than true. Although many of the AI search platforms do need work and will evolve over time, it feels like we are at a major turning point for human and robot interaction. Undoubtedly, there will be a learning curve.
Overall, as AI search engines evolve, people will start to interact with them in different ways. Like many of our grandparents had to learn how to use the computer when it was invented, and may still have a hard time with the technology, the following generations that grew up with iPhones and tablets know how to operate them and have come to rely on the technology of the day.
Huang equates past search engines to the computer instead of the iPhone, but says AI search is more full-service like iPhones are now. To operate a computer you often have to download files, install packages, and make constant security updates, he explains, whereas the iPhone does everything for you, and often updates in our sleep.
“I have a feeling AI-first engines will produce the iPhone effect,” says Huang. “Instead of searching for ‘best 5k training plan for beginning runners prior injury”, you’d say, ‘I’m just coming off a knee injury. I want to train for a 5k but don’t know where to begin. Where should I start? What should I read?’... The second is much, much more human.”
Read More: The Risks, Benefits And Ethics Of Artificial Intelligence
Carlee is a content writer and copywriter working in the Web3 space. Connect with her on Twitter @carlee_writes
This article and all the information in it does not constitute financial advice. If you don’t want to invest money or time in Web3, you don’t have to. As always: Do your own research.