What is the place of generative AI in the Internet of Things?

Generative artificial intelligence is a set of machine learning models that are created after being trained to guess the next set of words or the correct image based on a prompt.

Recently, the mainstream media has been very worried about Alexa, Siri and Google’s digital assistants. This is because, so far, these products have not used generative AI. Generative AI, is a set of machine learning models created that are trained to guess the next set of words or the correct image based on a prompt.

DALL-E, Stable Diffusion and Midjourney are popular for image generation and will be available in early 2021. But the latest hype is around large language models, specifically ChatGPT created by OpenAI. When someone enters tips and style suggestions, the results can be easily read and provide actual information or the illusion of information.

All devices are exciting and will have a major impact on the way work is done, content is created, and your business is built. However, there are concerns about the accuracy of generative AI. We still have to grapple with the ability to use generative AI for easy propaganda, fraud, and other malicious behavior, rather than simply blindly accepting generative AI as a world-changing technology.

That said, we are currently at the stage where technologists are blindly trusting generative AI, including investing billions of dollars in companies experimenting with use cases and new models. We're still at a stage where members of the media are spending hours trying to trick these models into misbehaving, or trying to prove that AI is sentient and potentially hostile to us.

But this is not the focus of this article. This article focuses on areas where generative AI will have a significant impact on how IoT is deployed and used. For example, where can we use it to improve user experience? What types of jobs can it help or take over? Besides words and images, what other generative AI models can help in the Internet of Things?

From Smart home begins. Rather than confusing Amazon Alexa with a chatbot, it and other digital assistants will continue to use natural language processing (NLP) to "understand" and then act on various task-based requests, such as "turn on the lights" or "good morning" ” to launch the wake-up call, while also adding a GPT-style chatbot to handle requests that require more in-depth communication.

A good digital assistant will not have just one or two models, but will consist of any version of the model to provide the most practical functionality to the user. There are also economic concerns. There can be fees for calling a chatbot, which requires a different business model, and not everyone is willing to pay for a subscription.

Additionally, we may soon see chatbot-style generative AI models being used in homes. Recently Home Assistant founder Paulus Schoutsen demonstrated how to use HomePod to access a GPT-style chatbot to tell stories to his children.

In fact, the utility of combining NLP, which is already part of digital assistants, with generative AI models is clear for SoundHound, which is introducing a platform that combines voice assistants with generative AI. So ChatGPT won't replace Alexa, but it may eventually become part of Alexa, with Alexa as the interface, and ChatGPT is just one of the many services it offers.

Other smart home areas where ChatGPT or generative AI models will have an impact include children’s toys, fitness services, and recipe or activity suggestions. That’s because generative AI is really just another reason to add connectivity and sentience to everyday objects, either providing personalized training data, or acting as a conduit for such services.

On the enterprise side, there is clear utility in using generative AI to help business people implement digital solutions without coding. One example is how Software AG is combining its Web Methods cloud-to-cloud integration platform with generative AI models to help employees figure out how to link data and various digital services. Ultimately, as more things become connected in buildings, production lines, commercial kitchens, and more, using simple written language to tell connected devices how to work with connected business software will help managers become more efficient and capable.

In industrial settings, the promise of ChatGPT comes with compelling use cases and considerations. Some support using generative AI for things like predictive maintenance. Generative AI models work by training on large amounts of data and then generating the most likely next element. So in a large language model, a generative AI model is being trained on a large amount of text and generates the next word or phrase that the model thinks is most likely to occur.

Presumably, with enough machine data, the model can decide what it should do next and send an alert if the expected result is incorrect. But honestly, this feels like overkill, as traditional anomaly detection is great for predictive maintenance and is much less expensive. Where generative AI might get interesting is by taking process data and suggesting alternative workflows, or by using written language to describe a workflow and letting the AI ​​code it for someone.

But there are also things to pay attention to. These models are only as good as their training data and may produce wrong answers in some cases, but can be written so well that it is difficult to determine whether they are wrong.

Given the intellectual property battles surrounding generative AI, the last concern is that “feeling” will become an issue. But in practice, it’s relatively simple to set limits on where the training data actually comes from—even if a model built on proprietary data is deployed outside the intended factory or enterprise.

Time and education on how generative AI models are created and how they work will address some of the IP issues. With only a few months into this cycle, I believe the future will see generative AI becoming as important and accepted as computer vision and NLP.

The above is the detailed content of What is the place of generative AI in the Internet of Things?. For more information, please follow other related articles on the PHP Chinese website!