David Tolfree, science and technology writer
There is an increasing number of news items that play on our fears about the use of artificial intelligence (AI). Many are based on the possibility of intelligent machines using data to replace human decision-making and eventually taking over control, leaving humans jobless and at their mercy. To a lesser extent, this is already happening in automated manufacturing processes that use robots but, for now, humans provide the controlling software by writing the algorithms for it. AI is the ability of a system or a program to think and learn from experience to simulate human behaviour and thinking.
Micro-nanochips are now pivotal to a nation’s economic growth since they provide the basic elements in almost all useful devices required to advance economic advancement and national security. They underpin the emergent technologies that use AI and big data analytics. The ownership of manufacturing facilities and supply chains is now central to political power in most advanced technological countries.
Alongside climate change, pandemics and nuclear war, the extended use of AI has become of concern to governments and is now among their priorities for discussion, but can it be justified as a threat to humanity? AI is simply an evolutionary development of computer and communication technologies that focuses on creating machines (robots) that can enhance and exceed the capabilities of humans. It is the recent developments in chip technology that enable vast amounts of data to be analysed and quickly acted upon that has increased AI’s capabilities.
AI is a driver of the 4th Industrial Revolution. It enables quick, accurate analysis from sources of data in almost every field of human endeavour. Operational medicine and healthcare are examples of the many beneficiaries of AI. For example, it has revolutionised pattern recognition from MRI scans, enabling immediate and more accurate data analysis for the early diagnosis of disease. Moreover, AI-powered robots featuring a range of precision sensors are assisting surgeons with complex operations and procedures that sometimes can be carried out remotely by experts who may not be available at the place required.
The Internet of Things (IoT) is benefitting from AI. It provides connections to networks for a variety of devices and has enabled smart homes and businesses. Virtual assistants such as Amazon’s Echo’s Alexa, Google Home, Apple’s HomePod and Samsung’s SmartThings are commercially available examples. These are the early ancestors of domestic robots. It is now possible to clone a robot to simulate actual humans both in physical form and voice. Done without consent, this is a possible unlawful threat to actors, performers and entertainers who may own the copyright of their work. AI can be used to enhance realism in computer-generated imagery (CGI), which is used in film production to create more immersive special effects. In addition, AI robots are being used in business to provide quick access to information and thus expedite decision-making.
The range of economic and social benefits of AI are presented in the UK AI Council AI roadmap, an independent report carried out by the AI Council in 2021 to provide recommendations to help the UK Government formulate a national AI strategy1. The reason for such a strategy is to deal with the volume of data produced, which globally is expected to grow from 33 zettabytes in 2018 to 181 zettabytes in 2025 (one zettabyte is a thousand billion gigabytes)2. The report emphasises the need to adapt the infrastructure to embed AI and data literacy so that its benefits, rewards and risks are understood by and acceptable to everyone. The three key areas outlined in the roadmap are to accelerate changes in R&D and innovation, skills and diversity, data infrastructure and public trust in supporting health, climate change and defence strategies.
AI technology presents challenges but also offers unparalleled enhancement and opportunities for humanity. Like all accessible technologies, it can be exploited for the wrong reasons, which is why governance and regulation in its use are urgently required. The current debate is about who and how AI should be regulated. Hi-tech companies and governments have too much self-interest to provide such regulation since it gives them immeasurable economic and political power. Independent bodies such as the UN have been suggested, but it would take a long time to obtain approval from all the member states and some would never agree to regulation anyway.
Foreseeing the problems of implementing regulation and the possible threat of AI, the European Parliament has taken the initiative of producing the first comprehensive law for regulating its use, the AI Act3. Talks with EU member states on the final shape of the law are in progress. The rules would ensure that AI systems developed and used in Europe are fully in line with EU laws and values. Similar discussions on regulatory issues are in progress in the UK, the USA and other countries.
As with all revolutionary developments that make radical changes to society, AI raises issues about the security and protection of information. AI software applications collectively known as ChatBots mimic human conversation through text or voice interactions online, enabling companies to create artificial conversations. They use a language model created from data and information. An example is Microsoft’s OpenAI’s ChatGPT (Chat Generative Pre-Trained Transformer). Google and Facebook have developed similar competing software systems. It and its variants can write and debug computer programs, compose music and teleplays, generate business ideas, write poetry and songs, translate and summarise text, and simulate entire chat rooms. OpenAI recently released ChatGPT-4, its latest version, which is rumoured to have 1.76 trillion parameters4. As humans and their activities are increasingly better simulated, we become more vulnerable to security threats. It is obvious, therefore, that regulation is necessary to control unlawful exploitation, although its global implementation will be difficult.
This brief comment has only outlined some of AI’s societal and industrial implications. In the 20th century, digital technologies replaced old methods and produced paradigm shifts in many industries. AI technology is driving innovation and will continue to influence the way we live and work. It may be some time before humans are replaced by AI robots, but with the exponential increase in knowledge and the need to analyse and exploit more data, they will become an essential part of our future society. Whether AI is a threat to humanity will ultimately depend on its development and how it is used by future generations.
David Tolfree, science and technology writer
References
1UK AI Council members: Smith, A. et al. (2021) UK AI Council AI roadmap [report]. UK AI Council.Available at: https://bit.ly/43Jr6zi
2Duarte, F. (2023) Amount of data created daily (2023) [press release]. Exploding Topics. April 3.Available at: https://bit.ly/3rPIB3R
3European Parliament (2023) EU AI Act: first regulation on artificial intelligence [press release]. European Parliament. June 14.Available at: https://bit.ly/43Sb4TJ
4Wikipedia (2023) GPT-4. Wikipedia.Available at: http://bit.ly/3OL0mZD