What is Generative AI?
Generative AI is a type of artificial intelligence that can produce content such as audio, text, code, video, images, and other data. Whereas traditional AI algorithms may be used to identify patterns within a training dataset and make predictions, generative AI uses machine learning algorithms to create outputs based on a training dataset.
Generative AI can produce outputs in the same medium in which it is prompted (e.g., text-to-text) or in a different medium from the given prompt (e.g., text-to-image, or image-to-video). Popular examples of generative AI include ChatGPT, Bard, DALL-E, MidJourney, and DeepMind.
- Generative AI, or artificial intelligence, is a form of machine learning that is able to produce text, video, images, and other types of content.
- ChatGPT, DALL-E, and Bard are examples of generative AI applications that produce text or images based on user-given prompts or dialogue.
- Generative AI is used in everything from creative to academic writing and translation; composing, dubbing, and sound editing; infographics, image editing, and architectural rendering; and in industries from automotive to media/entertainment to healthcare and scientific research.
- There are wide-ranging concerns about generative AI that touch upon legal, ethical, political, ecological, social, and economic issues.
How Does Generative AI Work?
Generative AI is a type of machine learning, which, at its core, works by training software models to make predictions based on data without the need for explicit programming.
Specifically, generative AI models are fed vast quantities of existing content to train the models to produce new content. They learn to identify underlying patterns in the dataset based on a probability distribution, and when given a prompt, create similar patterns (or outputs based on these patterns).
Part of the umbrella category of machine learning called deep learning, generative AI uses a neural network that allows it to handle more complex patterns than traditional machine learning. Inspired by the human brain, neural networks do not necessarily require human supervision or intervention to distinguish differences or patterns in the training data.
Generative AI can be run on a variety of models, which use different mechanisms to train the AI and create outputs. These include Generative Adversarial Networks (GANs), transformers, and Variational AutoEncoders (VAEs).
Generative AI Interfaces
Widespread AI applications have already changed the way that users interact with the world; for example, voice-activated AI now comes pre-installed on many phones, speakers, and other everyday technology.
Similarly, users can interact with generative AI through different software interfaces. This has been one of the key innovations in opening up access and driving usage of generative AI to a wider audience. Whereas early versions of generative AI required technical or data science knowledge to interact with the software, now, AI developers are designing user experiences in which prompts can be given and interactions can take place in plain language.
Here are some of the most popular recent examples of generative AI interfaces.
Created by OpenAI, ChatGPT is an example of text-to-text generative AI: essentially, an AI-powered chatbot trained to interact with users via natural language dialogue. Users can ask ChatGPT questions, engage in back-and-forth conversation, and prompt it to compose text in different styles or genres, such as poems, essays, stories, or recipes, among others.
Released in November 2022, a free version of ChatGPT is available for use online. OpenAI also sells the application programming interface (API) for ChatGPT, among other enterprise subscription and embedding options.
DALL-E is an example of text-to-image generative AI that was released in January 2021 by OpenAI. It uses a neural network that was trained on images with accompanying text descriptions. Users can input descriptive text and DALL-E will generate photo-realistic imagery based on the prompt. It can also create variations on the generated image in different styles and from different perspectives.
DALL-E can also edit images, whether by making changes within an image (known in the software as Inpainting) or by extending an image beyond its original proportions or boundaries (referred to as Outpainting).
Bard is a text-to-text generative AI interface based on Google’s large language model LaMDA (Language Model for Dialogue Applications). Like ChatGPT, Bard is a chatbot powered by AI technology that can answer questions or generate text based on user-given prompts. Google bills it as a “complementary experience to Google Search.”
In March 2023, Bard was released for public use in the U.S. and U.K., with plans to expand to more countries in more languages in the future. It made headlines in February 2023 after it shared incorrect information in a demo video, causing parent company Alphabet (GOOG, GOOGL) shares to plummet around 9% in the days following the announcement.
The History of Generative AI
Artificial intelligence has a surprisingly long history, with the concept of thinking machines traceable back to ancient Greece. Modern AI really kicked off in the 1950s, however, with Alan Turing’s research on machine thinking and his creation of the eponymous Turing test.
The first neural networks (a key piece of technology underlying generative AI) that were capable of being trained were invented in 1957 by Frank Rosenblatt, a psychologist at Cornell University.
Further development of neural networks led to their widespread use in AI throughout the 1980s and beyond. In 2014, a type of algorithm called a generative adversarial network (GAN) was created, enabling generative AI applications like images, video, and audio.
In 2023, the rise of large language models like ChatGPT is indicative of the explosion in popularity of generative AI as well as its range of applications.
How Is Generative AI Used?
Many generative AI systems are based on foundation models, which have the ability to perform multiple and open-ended tasks. When it comes to applications, the possibilities of generative AI are wide-ranging, and arguably, many have yet to be discovered, let alone implemented.
The ability for generative AI to work across types of media (text-to-image or audio-to-text, for example) has opened up many creative and lucrative possibilities. No doubt as businesses and industries continue to integrate this technology into their research and workflows, many more use cases will continue to emerge.
Current popular generative AI applications
Some examples of current use cases for existing generative AI models include:
- Creative, academic, and business writing
- Code writing
- Genetic sequencing
- Grammatical correction or analysis
Audio and speech models:
- Composing and songwriting
- Dictation and transcription
- Speech and voice recognition
- Sound editing
Visual and imagery models:
- 3D modelling
- Creative design
- Image editing
- Architectural rendering
Data generating models:
- Creating synthetic data on which to train AI models
Applications by industry
Industries are currently using generative AI in a variety of ways that will continue to expand as the technology and our understanding of it continues to develop. Examples of current applications across different fields include:
- Automotive industry: Synthetic data produced by AI can run simulations and train autonomous vehicles.
- Healthcare and scientific research: Scientists can use AI to model protein sequences, discover new molecules, or suggest new drugs compounds to test, while doctors and practitioners can use AI to analyze images to aid in diagnoses.
- Media and entertainment: AI can be used to quickly, easily, and more cheaply generate content, or (as a tool) to enhance the work of creatives like writers and designers.
- Climate science and meteorology: AI can simulate natural disasters, forecast the weather, and model different climate scenarios.
- Education: AI can be used to supplement classroom learning with one-to-one tutoring via a chatbot, or to create course materials, lesson plans, or online learning platforms.
- Government: The use of AI by the federal government varies by department. It has publicly released information regarding its use cases since 2022.
Of course, AI can be used in any industry to automate routine tasks such as minute-taking, documentation, coding, or editing, or to improve existing workflows alongside or within pre-existing software.
As generative AI models are also being packaged for custom business solutions, or developed in an open-source fashion, industries will continue to innovate and discover ways to take advantage of their possibilities.
One concern about generative AI is that algorithms can amplify or replicate existing discrimination and biases inherent in training data. Amazon, for example, created (and then abandoned) an AI-powered recruiting tool that was biased against women.
The Pros and Cons of Generative AI
Like any major technological development, generative AI opens up a world of potential, which has already been discussed above in detail, but there are also drawbacks to consider.
Overall advantages of generative AI include:
- Increasing productivity by automating or speeding up tasks
- Removing or lowering skill or time barriers for content generation and creative applications
- Enabling analysis or exploration of complex data
- Using it to create synthetic data on which to train and improve other AI systems
Disadvantages of generative AI include:
- Hallucination: This technical term refers to the tendency for certain AI models to generate nonsense or errors that do not correspond to fact or real-world or common-sense logic.
- Reliance on data labelling: Although many generative AI models can be trained in an unsupervised manner using unlabelled data, data quality and veracity remains an issue. Many tech companies, including OpenAI, Facebook, and Tiktok, rely on low-paid contract workers who perform data enrichment work such as labeling or generating training data.
- Difficulty with content moderation: Another concern is the ability for AI models to recognize and filter out inappropriate content. As is the case with data labeling, much of this work still relies on human contractors to tag and filter through large amounts of offensive and potentially traumatizing content.
- Ethical issues: In addition to labour concerns like the examples above, algorithms have been demonstrated to amplify or replicate existing discrimination and biases inherent in the training data. This can have resoundingly negative impacts. For example, Amazon created (and then abandoned) an AI-powered recruiting tool that was biased against women.
- Legal and regulatory issues: The legal system does not currently have a sufficient framework to address many of the implications of emerging AI technology. Some examples include:
- Copyright issues: Since generative AI models are trained on a vast quantity of data, it can be difficult to verify whether the materials included in the data or the resultant works generated are in violation of copyright laws.
- Privacy issues: Generative AI raises concerns around the collection, storage, use, and security of data, both personal and business-related.
- Autonomy and responsibility: AI technology raises concerns around liability. For example, when it comes to autonomous systems like self-driving cars, it is unclear how to determine liability in the case of accidents.
- Political implications: Generative AI raises issues around false or misleading information and the veracity of media such as photorealistic imagery or voice recordings. It can also interfere with processes that invite democratic engagement by falsifying a high volume of comments, submissions, or messages.
- Energy consumption: AI models have a large ecological impact, as they require a huge quantity of electricity to run. As the use of these technologies grows, so will the demand on the environment.
Which Industries Can Benefit from Generative AI?
Generative AI can benefit just about any type of field or business, by increasing productivity, automating tasks, enabling new forms of creation, facilitating deep analysis of complex datasets, or even creating synthetic data on which future AI models can train.
Generative AI is also widely used in many different government applications.
What Is the Concern Surrounding Generative AI?
As a new technology that is constantly changing, many existing regulatory and protective frameworks have not yet caught up to generative AI and its applications. A major concern is the ability to recognize or verify content that has been generated by AI rather than by a human being. Another concern, referred to as "technological singularity," is that AI will become sentient and surpass the intelligence of humans.
What Are Some Popular Examples of Generative AI?
Popular generative AI interfaces include ChatGPT, Bard, DALL-E, MidJourney, and DeepMind.
What Is Machine Learning?
Machine learning is the ability to train computer software to make predictions based on data. Generative AI uses machine learning algorithms.
What Is a Neural Network?
A neural network is a type of model, based on the human brain, that processes complex information and makes predictions. This technology allows generative AI to identify patterns in the training data and create new content.
The Bottom Line
Generative AI is an exciting new technology with potentially endless possibilities that will transform the way we live and work. Traditionally, AI has been the realm of data scientists, engineers, and experts, but now, the ability to prompt software in plain language and generate new content in a matter of seconds has opened up AI to a much broader user base.
As with any technology, however, there are wide-ranging concerns and issues to be cautious of when it comes to its applications. Many implications, ranging from legal, ethical, and political to ecological, social, and economic, have been and will continue to be raised as generative AI continues to be adopted and developed.