top of page
Avneet Kaur Galhotra

Artificial intelligence – The Evolution Has Just Begun


Computers have been one of the most empowering inventions given to the mankind and since its inception, their capability to perform various tasks has been growing exponentially. Continuous research is underway to increase the power of computer systems and such continuous research and curiosity of humans lead to development of a branch of Computer Science Technology named “Artificial Intelligence” that aims at creating computers or machines that are as intelligent as human beings.

Artificial intelligence and machines have become an integral part of our day to day life. Whether it is mobile devices, smart speakers, bots, digital assistants, smart instruments, drones, etc., AI has fundamentally changed the way we do things, plan our day, consume news, shop, communicate and interact with our family, friends and colleagues.

A huge amount of work and effort has already been put in this domain and has given some exceptional results like IBM’s super computer “Deep Blue” which won a chess match against the then world chess champion Gary Kasparov, in 1997 and Google’s Deep Mind AI program named “AlphaGo” that in 2016 defeated one of the best players, Lee Sedol in the Abstract Strategy Board game ‘GO’. However, this is just the beginning of revolution in AI, as this technology domain has a lot of potential to change the way we live and work today.

What does the term “Artificial Intelligence” actually mean?

The word ‘Artificial Intelligence’ is a combination of two words “Artificial” and “Intelligence”. The Oxford Dictionary defines ‘Artificial’ as made or produced by human beings rather than occurring naturally, especially as a copy of something natural while ‘Intelligence’ is defined as the ability to acquire and apply knowledge and skills. Thus, anything devised by human that has an ability to acquire and apply knowledge can be referred to as an implementation of AI.

There can be many definitions of AI, one being “A study about how to train computers to do things that humans can do better at present.” In general terms, Artificial Intelligence is a technology that can make machines do the unique things humans can do, like to talk, see, socialize, learn, and reason.

AI is basically accomplished by studying how human brain thinks, and how humans work, learn, and decide while solving a problem.

Early Days

The term “Artificial Intelligence” was coined by John McCarthy in 1956 in his first academic conference, The Dartmouth Conference of 1956. He defined AI as, “The science and engineering of making intelligent machines, especially intelligent computer programs.”.

Artificial Intelligence has been around since long. In the Greek myths, there are stories of mechanical men designed to mimic human behaviour.

The European computers were conceived as “logical machines” at a very early stage. Since then, there has been various major innovations in AI including development of Turing Tests for evaluation of intelligence by Alan Turing, development of ELIZA, a natural language processing computer program by Joseph Weizenbaum, development of the first computer-controlled autonomous vehicle named Stanford Cart, and development of interactive robot pets.

Application of Artificial Intelligence

The term AI has become ubiquitous in the technology industry and has numerous applications; few of them commonly seen in the following :

Gaming − Example chess, tic-tac-toe, poker, etc.

Machine Learning (ML) – Example Traffic predictions in navigation systems, pattern recognition, digital assistants (Siri, Alexa, Google Assistant, Cortana, Viki, etc.), video surveillance, chatbots, online fraud detection, etc.

Natural Language Processing (NLP) − Example Speech Recognition, Voice Recognition, Automatic voice output, Personal Digital Assistants, etc.

Robotics − Example Industrial robots for performing industrial activities, robots for house help, robots for performing surgeries, robots for help in movie industry, self-driving cars, etc.

Neural Networks – Example Pattern recognition systems like handwriting recognition, facial recognition, character recognition, handwritten character recognition, etc.

Vision Systems – Example a spying airplane, drone, clinical expert system, computer software used by Police to recognize criminals, etc.

Expert Systems − Example Flight-tracking systems, fire/burglary alarm systems, clinical systems, control systems, etc.

Fuzzy Logic Systems − Example Automobiles, consumer electronics, etc.

Data Mining − Example Big Data, Hadoop, Cloud Computing, etc.

Impact of AI on the Patent Industry

AI is seen to be largely impacting businesses world-wide. As per KPMG’s Venture Pulse Report, Venture Capital (VC) investment in artificial intelligence attracted $12B in 2017 as compared to approximately half in 2016, i.e., $6B.[1] Q2 of 2018 became the second record quarter with AI, total investments exceeding $2.3B, according to the latest PwC/CB Insights MoneyTree Report. [1]

International Data Corporation (IDC) has forecasted that investments on AI and ML will rapidly grow from $12B in 2017 to $57.6B by 2021, giving Compound Annual Growth Rate (CAGR) of 48%.[1]

The top 5 states (in 2018 ) in the United States for AI investments [3] are:

  • California with the investment of $1,917 million on 53 projects

  • Massachusetts with the investment of $247 million on 13 projects

  • New York with the investment of $110 million on 10 projects

  • Texas with the investment of $10 million on 3 projects

  • Washington with the investment of $9 million on 3 projects

On the IP Front, over 154,000 AI patents have been filed world-wide since 2010 with the majority of them being filed in the health sector (29.5%), industry-specific solutions (25.3%) and AI-based digital security (15.7%).[2] Out of these, 79,936 patents were filed in the United States only between 2010 and 2018. [2]

Interestingly, AI based marketing patents are becoming the fastest growing global category of patents with a CAGR of 29.3% between 2010 and 2018. [2] The other two fastest growing global AI patent categories are AI-based digital security with CAGR of 23.4% and AI-based mobility with CAGR of 23%. However, Machine learning (ML) dominates the AI patent landscape today, leading all the other categories of AI related patents. [2]

IAM mentions Microsoft as the global leader in ML patents with 2,075 number of patents. [1] EconSight patent analytics also indicate that Microsoft is leading the AI patent race with 697 high quality patents classified as significantly competitive by Microsoft itself. EconSight also mentions that Microsoft has created 20% of all the patents filed in AI world-wide by the top thirty companies and research institutions. [2]

Apart from Microsoft, tech giants like Google, Apple, Facebook and IBM have gone far ahead in applying AI to internet searching, classification of images, speech recognition and related fields. [4] Not behind in the race are, the tech companies like video streamer Netflix, payment processor PayPal, Salesforce and Facebook that embed AI tools in their products for better performance. [4] Technology companies like Infor, DeepMind (owned by Google), OpenAI and Banjo are also emerging fast as one of the major players in the industry.

Even Startups are racing to implement AI in robotics, smartphones, data centers, drones, etc. Crunchbase has listed over 5,000 startups that are majorly focusing on Machine Learning for the category of products, services and applications that they provide today. [1] AI startups had 2018 as their best funding year, with the investment of $99.5 billion which is approximately 10% of last year's total VC investment of $9.33 billion.[3]

Since 2013, VC investments in AI startups have been regularly going up for the next four years, showing a CAGR of about 36%.[3] However, the funding related to AI significantly jumped the last year, increasing by 72% as compared to 2017.[3]

With last year being the highest investment year in AI by startups, some huge competitions appeared in the AI, including the biotech firm Zymergen ($400 million), Dataminr ($392 million), Robotics Process Automation (RPA) company ($300 million), cybersecurity firm Tanium ($175 million), job site Zip Recruiter ($ 156 million), and UiPath ($153 million).[3] Nuro, a Silicon Valley startup that built a driverless delivery robot vehicle has raised nearly $1 billion from Softbank. Anobot, Cinnamon, CrowdStrike, Cylance, AutoLab AI, Pony.AI, C3 IoT, Nuro, and Tempus Labs are among other new AI startups in the race.

Further, Absentia VR, Niki.ai, Flutura, Uncanny Vision, Innefu Labs, Netradyne, Active.ai, FORMCEPT, Staqu, and Mad Street Den are few of the major emerging Indain Startups to watch out.

AI Patents and European Patent Office[5]

The main challenge for patent offices is the AI’s rapid growth across a range of technical fields. The EPO in its own words is well prepared to handle such patents, thanks to the principles developed in the case law of its boards of appeal and elaborated on in the Guidelines for Examination in the EPO.

The patent filing statistics at the EPO suggest that in AI related patents, a technical effect for algorithms is claimed, which is in line with the requirements laid down in the case laws of the EPO boards of appeal for patenting mathematical methods.

The source codes are protected by the copyrights and a mixture of copyright and patent protection is also seen in Europe.

When assessing patentability for mixed protections, the examiner applies two-hurdle approach where firstly, contribution of AI or ML steps or method to the technical character of the invention is judged with consideration being given to the nature of data used in the invention. Secondly, technical effect of the invention is examined with the specification being taken into consideration, including the test data.

Further, in Europe, to protect the fundamental algorithms related to the inventions, a compulsory cross-licensing is required to guarantee access to the algorithms, as mentioned in the Section 24(2) of the German Patent Act.

Since patents are considered as the best way to encourage innovation in the area of AI, there are a few topics which need a thoughtful discussion like increasing lifetime of AI patents due to the speed of evolution of AI technology and shortening the 18 month secrecy period until publication of the application due to possibility of unnecessary parallel research and development.

AI Patent Filing and stance of the Indian Patent Office[6]

NITI Aayog, policy think tank of the Government of India recently published a national strategy for Artificial Intelligence in June 2018. It laid down a roadmap for developing this sector in India.

According to this report, in order for India to ride the AI innovation wave, a robust intellectual property framework is required. Despite, number of government initiatives in strengthening the IP regime, challenges remain, especially in respect of applying stringent and narrowly focused patent laws to AI applications – given the unique nature of AI solution development. To tackle these issues, establishment of IP facilitation centers to help bridge the gap between practitioners and AI developers, and adequate training of IP granting authorities, judiciary and tribunals is suggested in the report.

Indian Examiners follow the Computer-related Inventions (CRIs) guidelines for examining AI-related inventions which excludes computer programmes or algorithms from patentability. Therefore claims directed to codes or algorithms is generally avoided by the applicants of AI-technology in India. However, to claim AI based software inventions in India, the applicant may rely on description of hardware, example computer system, server, sensor etc. along with AI algorithm in the claims. Claiming a method or a process of a device that uses AI is another approach followed by the applicants.

For example, IN239319 claims “A proactive user interface for a computational device having an operating system, the proactive user interface comprising: (a) an interface unit for … (b) at least one software application … (c) an artificial intelligence (Al) framework …” and IN228347 claims “A system that facilitates displaying objects comprising: an input component (110) …; a relationship component (120) …; and a display component (140) …”. Thus, claiming a device with hardware components or a device that uses AI.

AI Patent Filing and USPTO

Since there is no restriction in filing algorithms in the USPTO, US has emerged as one of the major countries with most number of AI patents. Out of 154,000 AI patents filed world-wide since 2010, 79,936 patents were filed in the United States between 2010 and 2018.[2]

At a recent event on Artificial Intelligence: Intellectual Property Considerations event held in January 2019, Director of the U.S. Patent and Trademark Office, Andrei Iancu mentioned that many AI efforts are underway at the USPTO itself, including the potential of engagement with academia and industry to help them identify the most advanced patent search tools. The USPTO has recently requested for Information from the industry to provide techniques that leverage AI for purposes of improving the patent examination processes at the USPTO.

Conclusion

In recent years, the emerging AI technology has provided some fruitful results like Duplex, a new AI assistant by Google that talks just like humans with “huh”, “uhm” where ever required, tiny elastic robots by Scientists at EPFL and ETH Zurich that are trained to change their shape based on their surroundings, a tendril-like soft robot with the ability to curl and climb around a plant stalk, a robot that can play Jenga, etc. With the emerging high-speed 5G networks, a powerful combination of high-speed 5G networks, the Internet of Things (IoT) and AI is being developed, termed as “Intelligent Connectivity” that will help build robots like never before.

America is at the forefront of the heart of this new world, i.e., “Intelligent Connectivity”. China is nowhere behind. In fact, while everywhere else there is a fear of AI taking over human’s job, China, the world leader in the digital economy, are the most optimistic about the impact of AI on people’s life and work. China has both, the tools and the knowledge, to take AI to the next level, however, uncertain or disruptive that may be.

This year, world will see few of the major AI innovations coming live. Volkswagen and Mobileye are going to launch a self-driving ride-hailing service in Israel. Volvo will bring its driverless truck trailers in limestone mines in Norway. GM will commercialize the Cruise AV that lacks steering wheel and pedals. Wing, a drone startup owned by Alphabet Inc, is going to launch a pilot delivery service in Finland. Airobotics, an Israel startup will survey construction and dredging of a new seaport in the Haifa city with its drones that can assess mine stock piles.

It can thus be safely said that with the advent of AI many dreams have come to life. It would be worth a wait to see how these super technologies change the way we live!

Reference List:

[1] 25 Machine Learning Startups To Watch In 2018

[2] Microsoft Leads The AI Patent Race Going Into 2019

[3] Venture Capital Funding For Artificial Intelligence Startups Hit Record High In 2018

[4] AI News: Artificial Intelligence Trends And Leading Stocks

[5] Patenting Artificial Intelligence

[6] Can Artificial Intelligence software be patented in India?

[7] What You Have To Fear From Artificial Intelligence


Featured Posts
Recent Posts
Archive
Search By Tags
Follow Us
  • Facebook Basic Square
  • Twitter Basic Square
  • Google+ Basic Square
bottom of page