Artificial Intelligence: Future and Trends to Watch

By Purshottam Purswani, Chief Technology Officer (CTO), Atos India

Artificial intelligence (AI) is posing a critical challenge to organizations. It is widely perceived that those who adopt AI successfully will gain tremendous advantages in terms of competitiveness and efficiency; the rest will be lagging behind.

However, AI is probably the most hyped of all technology trends – the magical and impossible is being promised, whilst the attainable and truly transformational is often being overlooked.

Future of AI:

The future of AI is not just about what we can do with it, but who will have access to control its power.

Could AI become ubiquitous, like electricity?

A technological power capable to drive economic and social change for the next generation?

As we peer into the near future, we see a move from ‘pioneering’ to the ‘application’ era of AI.

We are also seeing an increase in the use of AI in our day-to-day life, with AI being used for everything from increasing consumer protection against Fraud, Gaming, Natural Language Processing, Virtual Assistant, Conversation UX, Expert Systems, Vision Systems, Speech Recognition, Handwriting Recognition and Intelligent Robots to being at the heart of the Driverless Vehicle revolution.

With AI we will also get benefits in the Health Sector. Google’s ‘Deepmind’ has recently developed an AI system that detects eye disease quicker and just as efficiently as a world leading ophthalmologist.

In the healthcare vertical, we can already see a great demonstration of how AI could not just support, but also enhance the work of hospital staff by automating the transcribing of medical notes in healthcare, meaning crucial resources including doctors and nurses, can be deployed in other areas, saving money and making the system more efficient.

In addition, it is helping to clean up the planet by training AI algorithms to analyse drone footage and identify plastics and other foreign objects floating in seas and oceans which can then be cleaned.

Every day, AI is transforming our cities and our environment, making them cleaner, safer, and more efficient. With this sector developing so fast, these intelligent, autonomous, AI-driven systems and tools powered by data could also in the very near future, provide the answers to tackling some of societies biggest issues.

Atos is proactively working on AI and has developed following system for keeping the future in mind:

• AI base prescriptive Security Operation Centre (SOC) - detects all the signals left by attacks and alerts security managers about possible risk areas, even before the attack happens.
• AI base Digital twin (additional reference-1 and reference-2) using this manufacturer can create a real-time digital replica of all physical assets for comparing and analysing them in the future. Creates the ability to find new ways for improving production processes and allows engineers to create a prototype of a large or small product inside the digital world and undertake online testing, prior to physical testing.

The new breakthrough in AI that will disrupt the tech world in 2019:

Artificial Intelligence will have major impact in the automation, optimization and flexibility of connected ecosystems as we head towards 2019 & 2020. This will be especially the case in industrial supply chains as they become increasingly collaborative and responsive to market demands. 

In 2019, Artificial Intelligence and Machine learning will be disruptive technologies and will start to amplify some of the changes we will see in ways of Working.

Considering ways of working for industrial supply chains, connected Robots will be managed and controlled in completely new ways enabled by the offloading of selected processing to the Cloud. The same concept can be extended to include completely Connected Production Chains with Connected Production Assets.

There will be the potential with AI for remote control centres and technical operation centres that manage several production facilities (for example, drilling multiple oil or gas wells).

Most of Tech world companies will start using AI assistant and Conversational UX for ease of business and customer support.

Machine Learning and Artificial Intelligence will help to reach this level of automation and interaction, not just for releases, but also for proactive monitoring. The impact that this will have on traceability, duty and responsibility will be a key topic in the coming years.

On the software side, easy-to-use frameworks have been developed by companies like Google, Facebook or Microsoft.

These have been open sourced with a view to accelerate the diffusion of these technologies in research centres and industry. Thanks to this approach, DL has become one of the most active research topics in computer science, and has revitalized the research in Artificial Intelligence (AI).

Trends to watch:

Trends that change the world are Human centric-AI, it includes AI base decision making, by 2022 decision making will be the responsibility of hybrid teams of AI systems and people. AI will provide insights from huge volumes or streams of data, and humans will provide flexibility in cases of incompletely defined or unpredictable tasks, complex relationships with people, and cross-domain contexts. Humans will also take a lead on ethical aspects and cultural nuances, since satisfactorily embedding values and ethics into AI systems is still in its infancy. Similarly, creative AI will not be mature by 2022, so the spark of creativity and innovation will still mostly be people-driven.

Apart from decision making, another trend is Enterprise-scale AI systems, it will require the mastery of a range of technologies and company domains. This will demand the support of knowledge engineers who are well-versed in both business and domain specifics. Knowledge will need to be extracted from company sources (text, databases, machines etc) as well as from existing decision management systems (based on production rules), enriched with human interpretations, and marshalled into the AI services.

The technology and processes behind the AI enterprise will quickly evolve into an enterprise intelligence backbone, a kind of “artificial nervous system” pervading every part of the organization.

AI functions will follow the principle of subsidiarity of intelligence, where decisions are taken as close as possible to the point of action to avoid undesirable latencies. HPC services will externalize demanding ML tasks with the resulting models being deployed at the point of use.

The evolution of networks and computing capabilities will enable a complete spectrum of AI processing power from edge to cloud instances.

In the case of AI for industry, creating value from shared industrial data will only be possible through deep learning and advanced analytics. Such capabilities will be enhanced by High Performance Computing, which by 2022 will deliver exaflop performance.

We can expect enterprise AI to emerge from different areas and use cases. In some cases, it will replace existing systems, for example, with computer vision systems and associated intelligence replacing existing turnstiles, static access management, permission-granting or intrusion detection systems.

In other cases, AI will emerge alongside existing systems and blend with them, as in rule-based decision making or people management systems.

AI is likely to be initially introduced through proofs of concept and pilot programs, which will enable stakeholders to assess effectiveness and manage related challenges in response to a well-defined scope. Implementations will then be extended, sometimes rapidly as gains are attained, taking advantage of enterprise grade cloud services that scale.