If any technology has captured the collective imagination in 2023, it\u2019s generative AI \u2014 and businesses are beginning to ramp up hiring for what in some cases are very nascent gen AI skills, turning at times to contract workers to fill gaps, pursue pilots, and round out in-house AI project teams.\n\nAnalyzing the hiring behaviors of companies on its platform, freelance work marketplace Upwork has AI to be the fastest growing category for 2023, noting that posts for generative AI jobs increased more than 1000% in Q2 2023 compared to the end of 2022, and that related searches for AI saw a more than 1500% increase during the same time.\n\nThe recent AI boom has sparked plenty of conversations around its potential to eliminate jobs, but a survey of 1,400 US business leaders by the Upwork Research Institute found that 49% of hiring managers plan to hire more independent and full-time employees in response to the demand for AI skills. And 64% of C-suite respondents said they plan to hire more professionals across every job title because of generative AI technology. Of those surveyed, 59% also said that they are personally embracing generative AI in the workplace, with midsize companies taking the lead when it comes to generative AI adoption.\n\nSo what skills are companies hiring for exactly? Upwork analyzed data from its platform, examining the types of projects companies hired freelancers for thus far this year, to identify the most sought-after AI skills on the market. Here are the top 10 generative AI skills companies are seeking today, according to Upwork.\n\n1. ChatGPT\n\nAs evidence of its meteoric rise, ChatGPT was the most searched generative AI skill on Upwork in early 2023, just months after its launch at the end of November 2022. ChatGPT applications are widespread across industries, and organizations are interested in hiring professionals who have the skills to help them with using ChatGPT for content generation, task automation and scripting, translation, on-demand learning, technical support and troubleshooting, editing and proofreading, idea generation, calendar scheduling and management, and more. Understanding how to leverage ChatGPT in the workplace has quickly become an increasingly valuable skill that companies are interested in capitalizing on to achieve business goals.\n\n2. Natural language processing (NLP)\n\nNatural language processing (NLP) technology helps computers better understand human language to improve chatbots, AI assistants, automation, and other tasks. Language is constantly evolving, with nuances that can make AI-generated conversations feel unnatural, confusing, or robotic. NLP aims to create smoother experiences for those interacting with AI chatbots and other services that rely on generative AI to service clients and customers. These skills include expertise in areas such as text preprocessing, tokenization, topic modeling, stop word removal, text classification, keyword extraction, speech tagging, sentiment analysis, text generation, emotion analysis, language modeling, and much more. Most relevant roles for making use of NLP include data scientist, machine learning engineer, software engineer, data analyst, and software developer.\n\n3. TensorFlow\n\nDeveloped by Google as an open-source machine learning framework, TensorFlow is most used to build and train machine learning models and neural networks. As a popular and effective tool that assists companies with the development and deployment of AI models, the skill is in high demand across several industries and job roles. Relevant job roles include machine learning engineer, deep learning engineer, AI research scientist, NLP engineer, data scientists and analysts, AI product manager, AI consultant, AI systems architect, AI ethics and compliance analyst, among others.\n\n4. Image processing\n\nAI is being used to analyze and process images, while also pulling data and information from visuals and text documents, and interpreting or manipulating that data as needed. AI image processing enables organizations to analyze and extract data from documents such as invoices, purchase orders, packing lists, receipts, and more. It also has important applications in the healthcare industry, contributing to analyzing medical imaging from MRI and CT scans. There are several steps involved with image processing, including image acquisition, enhancement, restoration, processing, compression and decompression, morphological processing, image recognition, and data visualization. Relevant job roles include image processing engineer or scientist, robotics engineer, AI research scientist, quality control analyst, AR or VR developer, among others.\n\n5. PyTorch\n\nDeveloped by the Facebook AI Research (FAIR) team in 2017 as an open-source machine learning library, PyTorch is a popular framework that helps organizations build and train deep learning models and neural networks that can be employed in technology such as NLP and other applications. PyTorch is known in the deep learning and AI community as being a flexible, fast, and easy-to-use framework for building deep neural networks. Lauded features include dynamic computation graphics, a Python foundation, and automatic differentiation for creating and training deep neural networks. Relevant job roles include machine learning engineer, NLP engineer, AI research scientist, data scientist or analyst, medical imaging specialist, and AI ethics and compliance analyst. \n\n6. AI content creation\n\nGenerative AI\u2019s promise for content creation can\u2019t be denied, and more companies are turning to generative AI to create content such as blog posts, social media posts, graphics, articles, and even videos. However, generative AI isn\u2019t always successful at producing content that comes off natural or relevant to humans, resulting in clunky or awkward text that requires some level of human editing to make sense. As a result, there\u2019s been an increase in demand for professionals who can prompt gen AI engines effectively in the production of relevant content and cast an eye over AI-generated content to ensure it flows smoothly, makes sense, and will resonate with a human audience. Relevant job roles include AI copywriter, content strategist, AI content analyst, AI graphic designer, AI video producer and editor, AI chatbot content developer, and AI content compliance manager.\n\n7. Midjourney\n\nMidjourney is a generative AI service that was developed in 2022 to generate images using natural language prompts. It\u2019s currently available only for use through the company\u2019s official Discord server, using a bot that will generate four images per user request. Its release sparked some controversy, especially among members of the arts and graphic design communities, who worried that the technology could replace artists and human-made content. Concerns around copyright violation have also arisen \u2014 including lawsuits from artists who claim that the AI is infringing on artists rights due to its being trained on the existing work of millions of artists. However, the Midjourney research lab claims it wants to work with artists, and serve as a tool to help them create content easier. No matter where you stand on AI-generated art, the demand for Midjourney skills is high as organizations seek to capitalize on the technology to create custom ads, and as an alternative to searching Google Images for inspiration.\n\n8. AI chatbot\n\nAI chatbots have become commonplace in modern society, especially in e-commerce, customer service, and retail. Companies increasingly employ AI to reduce the workload on human representatives, typically rerouting customers to helpful documentation, direct answers to common questions, and support for simpler issues before connecting them with a human representative. Using AI chatbots, organizations can better streamline the customer experience, but these services also require regular maintenance and management by human workers. Organizations are looking for professionals who can test and debug, deploy and integrate, and analyze and monitor chatbot services. They\u2019re also seeking skills around APIs, deep learning, machine learning, natural language processing, dialog management, and text preprocessing.\n\n9. Model tuning\n\nModel tuning is the process of identifying and establishing the ideal settings and parameters for machine learning and deep learning models. By adjusting and fine-tuning these settings, teams can improve the performance and efficiency of their machine learning models. Model tuning uses trainable parameters, which are learned internally from data, and hyper parameters, which are configured by the user, to ensure the model generates the most accurate outcomes possible. With generative AI, this skill is important for creating quality consumer-facing products and services. Relevant roles include machine learning engineer, data engineer, deep learning engineer, data scientist, AI research scientist, quantitative analyst, AI consultant, and data analyst.\n\n10. Stable Diffusion\n\nStable Diffusion is a deep learning model that produces high-quality artwork and images based off complex and detailed user prompts. It\u2019s designed to continuously learn from user inputs and improve its outputs over time. With Stable Diffusion, users can also opt to modify and edit existing images to remove objects, crop and adjust images, and change the colors of objects or subjects. As with Midjourney, artists and designers have expressed concern over the technology, pointing to the same copyright issues, including claims that the software has removed data from copyrighted works without crediting creators. But Stable Diffusion remains an in-demand AI skill that organizations seek to help create custom advertisements and other content such as product images and social media content.