Identifying the various forms of generative AI

- You've likely heard of ChatGPT, but have you heard of Claude, Midjourney, VoiceBox, ElevenLabs, Synthesia, HeyGen? The gen AI space is rapidly evolving and while most people are familiar with solutions like ChatGPT that generate text, not as many people are aware of other forms of gen AI like image generation, music generation, voice generation, and cloning. There's also video generation that can be used within talent acquisition and new solutions seem to pop up almost daily. While ChatGPT and other large language models, otherwise known as LLMS, are known for their ability to generate human-level text output in nearly a hundred languages, they're equally adept at understanding natural language and data sets, including numbers; this is because they've been trained on massive amounts of content and through the use of a specific type of neural network known as a transformer which is the T in ChatGPT. LLMs can read, understand, and analyze text-based data. While ChatGPT might be the most widely known LLM, there are literally hundreds of them in existence from publicly available models like Anthropic's Claude, Google's LaMDA, which powers their Bard AI chat service, and Meta's LLaMA, to private or custom LLMs developed by HR tech providers or even your own IT staff. The custom-built LLMs specifically designed for HR-related use can perform better than general models for certain tasks. Beyond text understanding, generation, and analysis, there are gen AI solutions such as Midjourney and Stable Diffusion that can turn simple text into stunning images. In some cases, these solutions are capable of creating photorealistic images that can essentially eliminate the need for stock imagery because you can literally create anything you can imagine with a simple text prompt. Text-to-speech generation solutions like VoiceBox and ElevenLabs allow you to create audio output that's virtually indistinguishable from real people and they allow you to choose from many different voices and languages. Some solutions will even allow you to clone a specific person's voice and then create audio output that sounds almost exactly like the real person from any simple text input. Going one step further, there are solutions available like Synthesia and HeyGen that can generate video output of near-lifelike human avatars complete with voiceovers in over 120 languages. With this technology, you can create videos as easily as you would a basic PowerPoint slide deck, and you can add AI-generated music using solutions like Meta's AudioCraft as simply as typing what you'd like to hear. It's important to understand that gen AI is so much more than ChatGPT. Be sure to think of gen AI in all of its forms when it comes to looking at your talent acquisition processes and how gen AI can be used to drive efficiencies, improve experiences, and lead to improved outcomes for applicants, recruiters, and hiring managers.