By Professor Louis C H Fourie
Humans are pretty adaptable – well in most cases. Especially in the innovative and fast changing world of technology, humans proved themselves to be very adaptable. Just think of the various industrial revolutions and how we adapted from an agrarian to an urban and mechanised life; electricity and mass production; computers, the internet, automation and nuclear energy; artificial intelligence (AI), Internet of things (IoT), and robotics; and eventually human-machine collaboration.
Even on a more personal level the quick integration into the daily lives of people of the telephone, motor car, radio, television, personal computers, mobile phones, the internet, and smart phones, is remarkable. Interesting is that the pace of the adoption is accelerating – from 46 years for electricity to achieve a penetration rate of 25%; 35 years for the telephone, 16 years for the personal computer, and only five years for the smart phone.
More recently, the popular artificial intelligence chatbot from OpenAI, ChatGPT, has set new records in adoption. ChatGPT that can generate articles, essays, jokes and even poetry in response to prompts, acquired 1 million users just five days after launching on November 30, 2022. This makes it the fastest-growing consumer application in history. It took Instagram approximately 2.5 months to reach 1 million downloads, whereas Netflix took 3.5 years to reach 1 million users.
By January 2023, ChatGPT is estimated to have reached more than 100 million monthly active users, a mere two months after its launch. It took TikTok about nine months after its global launch to reach 100 million users and Instagram about 2.5 years, according to Sensor Tower. According to UBS analysts, it is the fastest rise in a consumer internet application ever.
Suddenly, the way people interacted with the digital world completely changed. Computer coders, office workers and students flocked to generative AI to increase their productivity and ask questions on a wide range of topics, while regulators around the world scrambled to understand the technology and make new laws to ensure it will not be used to harm people.
People are increasingly relying on the capabilities of AI, and are seamlessly incorporating them into everyday routines. Generative AI are also becoming add-ons and plugins to word processors and spreadsheet software (e.g. the Microsoft Office and PowerBI Copilot) and search engines (e.g. Bing).
This rapid adoption of generative AI leaves us with a remaining question: What will be the next innovation in AI that will capture our imagination? Some tech companies are working frantically on promising next-level Generative Pre-trained Transformers (GPTs), humanoid robots, AI lawyers, and AI-driven research.
The next-level AI appears to be ready from a technological point of view, but whether they satisfy the three most important criteria for success is not certain. For AI to have a lasting impact, it needs to be not only technologically feasible, but also economically viable, and normatively acceptable – that means it has to comply with the values that society demands all should conform to. Let us look at some future AI technologies.
Multimodal AI systems
Multimodal chatbot systems based on the latest LLM technology that can handle different types of data, such as images and speech, as well as text, will soon be widely available.
Auto-GPT, an advanced AI tool by Significant Gravitas, is making waves in the tech industry. When Auto-GPT is given a general task, for example the planning of an anniversary party, it divides the task into sub-tasks and completes it without human intervention. Auto-GPT makes decisions according to predetermined rules and goals and could be used in a variety of applications.
AI legal support
AI offers promising support to the legal profession. Recently it was reported that a new startup company, DoNotPay, has built a legal chatbot based on LLM technology. The company claims that their chatbot can advise defendants in court via an earpiece. The AI listens to the court proceedings and then in real-time provide legal arguments into the ear of the defendant, who then repeats them verbally to the magistrate or judge.
The company offered the system to two defendants who had to appear in court for speeding tickets. However, the company drew heavy criticism and was also taken to court for practising law without a license. DoNotPay thus postponed their AI’s debut in court.
It seems that the success of this legal support technology will not be decided by technological or economic constraints, but by the rules and authority of the legal system.
However, the costs of litigation are high and many people can often not afford representation. The economic potential of automation in the legal profession is thus vast and robots might possibly in future represent defendants in court despite current opposition from the legal system.
Humanoid robots
In the past few years humanoid robots – robots that look and move like humans – have significantly advanced. Several companies are now developing humanoids that can mimic human behaviour and complete complex tasks. Many of these humanoid robots are being used in warehouses and factories.
Advancements in computer vision, as well as power-dense batteries, enabled robots to navigate complex environments. Several companies, such as Figure AI, 1X, Apptronik and Tesla have in invested in humanoid robots due to their advantages over other robots. Humanoid robots can navigate, manoeuvre, and adapt much easier since their operating environments have been built around human needs.
Research support
Scientists are more and more using AI and machine learning to discover patterns and insights in research data. Machine learning entails the use and development of computer systems that are able to learn and adapt over time without following explicit instructions.
Using these technologies researchers often employ generative AI systems to assist in the formulation of hypotheses. For instance, researchers from the University of Liverpool used a neural network (a computer system modelled on the human brain and nervous system) to rank chemical combinations for battery materials, to guide their experiments and save valuable time.
While generative AI cannot presently formulate hypotheses independently, it can assist researchers to approach problems from new perspectives and can lead to unexpected discoveries.
Similarly, generative AI can be used in businesses to understand company data by discovering patterns and trends in the data.
The future
All around the world, rapidly evolving AI is becoming a fundamental part of everyday life. According to Kurzweil, technological progress is exponential and not linear.
Organisations that fail to keep up with the speed of technological change risk falling behind and losing their competitive edge. Fast technology adoption is no longer a luxury; it has become essential for continued success.
But although the progress in AI is remarkable, the success does not depend only on technological innovation and computing power. There are also economic viability and the ethical aspects and values of society that must be weighed carefully in each case. We will probably see many battles in future around the ownership of intellectual property.
To a certain extent the future of AI remains an accelerating race into the unknown. It is a race that offers immense opportunities for solving the problems of the world, but it also carries with it certain risks.
Professor Louis C H Fourie is an Extraordinary Professor in Information Systems at the University of the Western Cape.
BUSINESS REPORT