By Hiba Hassan, head of design and visual communications, SAE
On November 30, 2022, OpenAI, a San Francisco-based company, released Generative Pre-Trained Transformer (ChatGPT), an AI-based chatbot. ChatGPT is based on OpenAI’s GPT-3 AI platform. It uses a blend of supervised and unsupervised learning to operate some of the biggest language models in the world. It can recall previous conversations and eliminate inappropriate or harmful responses. It also provides answers from a vast amount of data from the internet, and it is currently available for free to the public. In its first five days, the chatbot had already attracted over 1 million users, with positive reviews from influential publications such as The New York Times, which referred to it as ‘the world’s best AI chatbot’, and The Guardian, which said it generates ‘human-like text’. Despite its capabilities, ChatGPT has been reported to need assistance with data bias in its training.
It is easy to envision AI models knowing everything all the time, particularly with the vast amount of online knowledge. However, the risks of consuming information online must be taken into account, as much of it needs to be more accurate. You.com, a search engine launched last year, began offering a ChatGPT-style chatbot bringing more AI technology to the wider web. The search engine hopes to stand out by providing answers to more-current questions, but the chatbot has been criticized for confidently publishing incorrect answers. A disclaimer on the website states that the product is still in beta and may not always be accurate, and You.com is not liable for any inaccuracies in the generated content.
The surge in popularity of ChatGPT has caused worry, as Microsoft’s January announcement to include the technology in their search engine, Bing, means Google will have their first real competitor. LaMDA, a similar technology from Google, has yet to be released due to potential issues such as providing embarrassing answers or spreading hate speech. This has highlighted the need for accurate data to train AI systems and for governments to introduce regulations to guarantee the truthfulness of the information. In response, China has declared legal measures on AI-generated content, requiring government approval before its release, attaching identities to accounts, and labeling AI content.
The Technology Innovation Institute (TII) in Abu Dhabi is playing an important role in AI and Natural Language Processing (NLP) by having developed the world’s most comprehensive NLP model for the Arabic language, Noor. NLP is a key component of the artificial intelligence industry, allowing computers to comprehend spoken and written language in order to develop language translation tools, chatbots, and virtual assistants such as Siri and Alexa. The Noor model could potentially give the Arab world an advantage in the digital revolution, as tools like chatbots, market intelligence, and machine translation currently favor English and Chinese-speaking markets. The model is intended to be used by companies and academics to create new tools, such as sentiment analysis across social media, and to create new Arabic virtual assistants. Additionally, Dr. Ray Johnson, Chief Executive of TII, has declared that the public will also have access to the Noor model as an open-source model.
Regardless of their limitations, these models are part of a shift in AI being used to create new forms of art, music, writing, and code. It gives a window to a potential future. Still, it raises questions about the uniqueness of AI-generated content and several societal and ethical questions, especially the ability to replace knowledge-based jobs like customer service and content creation.
NLP chatbots have been used for marketing for some time now, but with AI power comes even more potential. For example, Copy.ai can generate personalized content for marketing campaigns, tailored emails, social media posts, and blog posts in mere seconds. Furthermore, dedicated ChatGPT chatbots could, in theory, take the place of customer service jobs, providing accurate information about products or services quickly and with the ability to customize chat responses according to individual customers’ interests and preferences. Additionally, ChatGPT can analyze customer feedback and uncover patterns and trends that can be used to inform marketing strategy and decision-making.
In conclusion, AI language models like ChatGPT hold great potential for various industries but also raise important ethical considerations. As McKinsey Global Institute predicts that intelligent agents and robots will replace 30 per cent of human labor by 2030, governments and organizations must address the potential displacement of jobs by investing in reskilling and retraining programs. The next decade will be focused on efficiency, with companies that can save time and operate faster through the integration of AI models like ChatGPT will have a competitive edge. Therefore, it’s imperative for organizations to closely monitor the development and use of these models to ensure responsible implementation and mitigate potential risks while also reaping the benefits of efficiency and productivity.
