Artificial intelligence tools are already ubiquitous in civilian and security applications. Soon, the rapidly developing AI technologies will wash like a tsunami over many fields.
In a nutshell
- AI-powered systems increasingly serve as assistants to humans
- Many AI machines will also partner up with humans in some areas
- AI breakthroughs will impact superpower rivalries
While authors of dystopian novels and sci-fi films often depict artificial intelligence (AI) as subjugating humanity upon reaching a critical threshold, the technology is more likely to assist and enhance human activity. The adoption of AI technologies – computer systems able to perform tasks that typically require human intelligence – has already started. The process will continue incrementally, often in ways invisible to consumers. The proliferation of these technologies will be ubiquitous throughout the world in the near term.
The current generation AI
The ability of computers to evaluate information, make choices and act on decisions has made remarkable progress over the past decade. What network operators need to pay attention to next is a particular method of AI, called machine learning (ML).
Machine learning processes try to mimic the functioning of the human brain, taking clues from how brain cells work in a neural network. This approach is data-driven, providing inputs that become the basis for establishing cause-and-effect relationships in a manner similar to how human brains create knowledge and make judgments.
While machine learning may not lead to the most advanced forms of AI in the future, its impact on contemporary developments in the field is unquestioned. Machine learning will, without a doubt, be part of a group of technologies that bring about a new generation of computer services. Pairing computers that can make better decisions with sensors – such as the Internet of Things (IoT) devices that can collect more and richer information – will lead to new capabilities, synchronizing the benefits of advancements in both technology fields. In short, machine learning technologies will likely have the most impact on networkers before very long.
In the next five years, ML-enabled technologies that can deliver reliable, scalable, cost-effective capabilities are going to wash over the marketplace like a tsunami in many fields in the private and government sectors. Today, machine learning capabilities are common in many widely deployed technologies, including speech and facial recognition. When the final numbers were tallied, one market research report estimated that total business worldwide, “including software, hardware, and services, are expected to total $156.5 billion in 2020.” This value would represent more than a 12 percent increase from the previous year – remarkable growth, given the drag on the global economy from the Covid pandemic.
AI and humans
Where ML technologies will have the most significant influence will likely be where these technologies complement rather than supplant human work, blending the activities of machines and humans. The relationships between AI and human networkers will be multifaceted. Thomas Malone, director of the MIT Center for Collective Intelligence, posits that three types of human-machine collaboration will dominate:
- Machines serve as assistants. Such a relationship is already occurring; computer-assisted surgery is an obvious example
- Machines function as peer partners with humans. One team of researchers postulates that “[b]y combining the two concepts of assistant systems and cognitive architectures, we can create a system which is capable of seamless human-machine interaction and integration, like a peer to [its] user instead of a servant or a simple assistant”
- Machines function as managers, directors, and supervisors. Machine learning technologies are already employed in reviewing and making recommendations on human performance, such as examining prescriptions issued to pharmacists.
The areas where the capacity of AI to process volumes of data influences networks are many, and they are expanding rapidly. For example, AI is already employed in defense networks, providing layers of cybersecurity protection. ML technologies are helping identify cyber threats and direct countermeasures. All these methods heavily depend on automated processes involving little or no human intervention. The proliferation of machine learning-enabled technologies in this field is likely to continue expanding. The trend of networkers operating in a distributed fashion remotely on various digital devices and systems will further accelerate the demand for “smarter,” scalable, and proliferated AI-empowered cybersecurity.
AI’s most significant importance will be in offering a great leap in the capacity of computers to create knowledge from data and do something with it.
Coupled with technologies that allow better gathering of raw data – everything from laser scanners at supermarket checkouts to crewless aerial vehicles on the battlefield – the volume of data available has increased dramatically. According to a projection by Statista, 74 zettabytes of data was created in 2021 (a zettabyte is a trillion gigabytes). That is up from 59 zettabytes in 2020 and 41 zettabytes in 2019. There is going to be a great deal more.
The increases are so exponential that 90 percent of the world’s existing data has been generated in the last few years. And this exploding growth will increase further. Among its drivers is the deployment of 5G telecom systems: the fifth generation technology standard for broadband cellular networks massively expands the capacity to move data and bring computer processing power closer to the data source. Other drivers include the relentless expansion of cloud services (storage and computing power provided as service distributed over multiple data centers used for storing information) and the technologies that will follow.
Any technologies that help networks access this data are going to provide a crucial competitive advantage. The intermingling of IoT and AI will enhance data management. Machine learning is already widely applied to software for IoT devices and services to make them more innovative, secure and productive. In turn, since machine learning benefits from large volumes of data to operate successfully and improve performance, and networks of IoT sensors and devices provide an enormous amount of information. The synergy of the two create added value to the business, industrial, consumer, and government applications.
AI, however, will not solve every problem in networking competition for one simple reason: to work best, AI needs a prodigious quantity of data. Unfortunately, not every problem that needs solving is rich in data. Machine learning technologies are effective when they can identify established patterns in bounded environments. A good example is traffic systems, where computers could learn from past commuter behavior to manage future traffic flows. However, other challenges can be highly complex and chaotic, and often involve impactful activities with minimal data sets. The 9/11 attacks presented such a situation. Also, everything the White House knew at the height of the Cuban Missile Crisis could have fit in a binder. Sometimes the deadliest challenges are the most data-poor.
Challenges brought by IT
As with other advanced technologies, AI raises concerns about the abuse of privacy and other civil liberties. One example is public anxiety over utilizing AI-based facial recognition technology by police departments. Several leading tech vendors, including Microsoft, IBM and Amazon, announced that they would limit sales of such software to law enforcement agencies. Another concern is the use of ML tools to create elaborate, “deepfake” misinformation campaigns.
Regulatory structures will also significantly constrain or speed up the deployment of new AI technologies. There is no extant comprehensive regulatory framework for AI-enabled technologies. However, once regulatory frameworks for those technologies emerge, they could crossover into labor, environmental, safety, security, privacy and many other regulatory systems.
Regulations that govern the internet were hammered out under the laws pertaining to commercial telecommunications networks. Unlike the worldwide web, AI is a new-generation hybrid capability that cannot be treated like telecommunications technologies. Managing this capability will require a new approach. Misapplication of regulatory structures could undermine innovation, limit competitiveness and inhibit economic growth.
Levels of public-private collaboration in AI development will also be critical. As noted previously, no aspect of public-private partnership on emerging technologies has greater import than how the U.S. will engage in the great-power competition, particularly with China. China is a potent competitor in this sphere. Working in its favor is the large size of its economy and its capacity to exploit technology. Moreover, China’s practice of civil-military fusion allows the government to seamlessly integrate all elements of national power, including industries and companies, to press the developments of its AI capabilities. Beijing has identified AI as a critical strategic development. The U.S. and other liberal states will not keep pace without efficacious cooperation between the public and private sectors.
The most likely scenario is that machine learning technologies will be rapidly and widely integrated into public systems, commercial enterprises and consumer products throughout the world. They will impact virtually every field of human endeavor. They will change the nature of work, but rather than supplant humans, the most common model will involve various forms of human-computer collaboration. Competition between China and the liberal states will largely fuel and shape the pace of AI adoption and proliferation.