Intro to Generative AI

Generative AIs are a type of artificial intelligence that uses machine learning techniques to create new content or predictions, such as text, images, audio, or video based on patterns they learn from input data.

Generative AIs are trained by feeding them massive amounts of data, such as text, images, or sound. These inputs are then studied to recognize patterns and rules, which are used to create new, original content in a similar style or format.

Generative text AIs can be used to:

• Write and critique content, such as articles and blog posts
• Automate interactions, such as emails, texts, and chatbots
• Personalize tutoring by answering specific questions on content
• Translate text from other language
• Organize and assist in routine tasks as a virtual assitant
• Analyze and report on inputted data
• Write and edit program code

Generative image AIs can be used to:

• Create images both new and based on inputted templates
• Customize designs, such as clothing or housing proportions and colors before manufacturing and purchase
• Procedurally generate textures, terrain, and characters for videogames
• Augment training by creating examples images of certain scenarios or diagnoses
• Clone human faces and apply them to other images Generative image AIs have risks and limitations, such as being used to mislead others, unconsensual use of other people’s likeness, and generating biased or offensive content.

Generative audio AIs can be used to:

• Create music in a variety of styles
• Convert inputted text to speech
• Clone voices for use in dubbing film or voiceovers
• Restore and enhance noisy or old recordings
• Enhance customer support by providing more interactive audio support

Generative videos AIs can be used to:

• Very accurately replace the appearance of one person with the likeness of others
• Automate labor-intensive video editing and special effects tasks
• Synthesize training data for applications such as autonomous vehicles, security, and medical training
• Generate analytic overlays for sporting events
• Enhance low quality or damaged videos
• Dynamically generate content specific to the user

Generative text AIs have limitations, such as generating incorrect or nonsensical responses, not understanding context, and not having access to real-time or very recent data. They may also generate offensive or biased content.

Generative audio AIs have risks and limitations, such as creating content without proper licensing permissions, being used to impersonate and mislead others, and feeling impersonal when used for tasks such as customer support.

Generative video AIs have risks and limitations, as being used to mislead others, unconsensual use of other people’s likeness, and generating biased or offensive content.

There are ethical concerns when using generative AI, such as creating misleading content to spread misinformation, issues with copyright and originality, and issues related to the use of biased and private training data.