Friday, May 8, 2020

Artificial Intelligence and 5G


Why is 5G Relevant to the Field of AI?
My focus this week has been exploring the intersection of artificial intelligence and 5G. When looking at the potential risk and possibilities with this combination and perhaps talking to industry experts about their thoughts. Who is studying the effects or changes that this is leading to? There is a big investment by telecommunication companies into these areas, and there is an expressed fear by many in regards to the possible adverse consequences. It may change the way that we use applications within the field of artificial intelligence, that you can do more with a small device because your connections allow for data to move more swiftly. Thus on an interface many applications will be decentralized with cloud (on a server). It is clear that if good things can be extended then less good aspects can too.
I will explain and go through:
How 5G works, Spectrum allocation, Robotics and 5G. EU’s Security Policy relating to 5G.China’s 5G expansion. Google’s purchase of Fitbit for $2.1 Billion.
1. How Does 5G Work?
5G networks are digital cellular networks, in which the service area covered by providers is divided into small geographical areas called cells.
A good explanation has been attempted in Digital Trends.
The article explains spectrums: low-band, mid-band and high-band.
“Low-band spectrum can also be described as sub 1GHz spectrum. It is primarily the spectrum band used by carriers in the U.S. for LTE, and is quickly becoming depleted. While low-band spectrum offers great coverage area and penetration, there is a big drawback: Peak data speeds will top out around 100Mbps.
Mid-band spectrum provides faster coverage and lower latency than you’ll find on low-band. It does, however, fail to penetrate buildings as well as low-band spectrum. Expect peak speeds up to 1Gbps on mid-band spectrum.
High-band spectrum is what most people think of when they think of 5G. It is often referred to as mmWave. High-band spectrum can offer peak speeds up to 10Gbps and has very low latency. The major drawback of high-band is that it has low coverage area and building penetration is poor.”
There are a lot of hairy displays of 5G in fancy connective points.
This has additionally been displayed in a manner of evolutionary progression, as is often done within technology.
This becoming the first city with tech coverage of 5G becomes a goal in and of itself. In one sense an enabler of technology.
2. Artificial Intelligence, 5G and Spectrum Allocation
On the 25th of October Karen Hao from MIT Technology Review wrote an article called: “DARPA is betting on AI to bring the next generation of wireless devices online.” This article placed its focus on an important aspect of wireless communication. Because there is a planning of how the bandwidth is supposed to work via the radio spectrum. The electromagnetic waves in the frequency range of the radio spectrum is called radio waves — and these are widely used in modern technology, especially within telecommunication. To hinder interference between different user’s radio waves is strictly regulated by national laws, coordinated by an international body, the International Telecommunication Union (ITU). Karen Hao’s article discussed a recent hackathon organized by the Defence Advanced Research Projects Agency (DARPA). In the article she writes:
“The challenge was prompted by the concern that the growing use of wireless technologies risks overcrowding the airwaves our devices use to talk to one another.”
Since the frequencies are divided up between different commercial and government entities those who owns rights to a bit of spectrum rarely uses all of it 100% of the time. The new 5G networks that are emerging makes this more urgent as it is projected to grow enormously over time
Graph displayed in MIT Technology Review retrieved the 28th of October 2019
The challenge that were to be tackled in the competition issued by DARPA was to design a new type of communication device that doesn’t broadcast on the same frequency every time. The suggested proposal was to use machine-learning algorithms to find the frequencies available. They are wanting to move from a system managed through ‘pen and paper’ to a system managed by ‘machines autonomously’.
Karen Hao describes the final where 10 finalists competed against each other. This included scenarios in supporting communications in a military mission, an emergency response and concert venue. A team from the University of Florida took home the $2 million grand prize.
2.1 IoT and AI Revolution
To see the most accentuated possibilities that lie within 5G it may be interesting to look at a few of the actors that have much to gain from its expansion such as Samsung. In a news article written by Samsung on the 28th of October 2019 there is a description of the IoT and AI Revolution.
This article claims that 5G connectivity will allow different solutions to access more data real-time which will make solutions very different. 5G uses ‘far less power’, leveraging IoT sensors that can last for year. Within farming adjusting irrigation levels and in manufacturing within factories placed to detect when the condition of equipment changes.
There is in addition to this the possibility to carve out multiple uses for doctors to secure handling of patient data. The article states:
“5G is up to 20 times faster than 4G. But this latest generation of networks is hardly just ‘faster 4G.’ Coupled with record-setting low latency, 5G’s speeds will allow developers to create applications that make the most of improved reaction times, such as near real-time video transmission for sporting events or security purposes.”
2.2 But is it a Race?
As pointed out earlier this year there has been a great effort by the companies wanting to expand 5G connection to frame the competition as a race. Here is an entire microsite from industry lobbying group CTIA titled “The Race to 5G.” There are serious trade-offs in the spectrum and they are being neglected due to the speed the deployment is being made.
“And once the ‘race’ to build out 5G in big cities is ‘won,’ the pressure to expand access to other places in the country will vanish, making that divide even worse. It is worth carefully considering all of these things before giving in to haste.”
Therefore we should not get too carried away with this ‘race’ and be blind to possible adverse consequences that this expansion may bring.
3. Artificial Intelligence and 5G Robotics
3.1 The Speed that May Enable Further Usage of Robotics and AI
If we disregard the environmental damage that AI and Robotics could cause considering the heavy resource requirement it has, let us think about what it actually could do. Phone companies and telecommunications tout the amazing new things you can do with your phone should you be able to ‘up your speed’ or bring it to the next level. However, truth be told what would you do? Not much more than before. Streaming is already working rather well, and you can play candy crush etc. Yet we did not imagine what we would be doing with this type of speed or connectivity before it arrived. Everyone, or most people, sitting with their eyes glued to a phone on public transportation. The public spaces filled with people that are somehow cluelessly walking with their phone stretched out in front of them to navigate, read, play or socialize on the go. Beyond your immediate use there are several use cases in industry we should be aware of.
3.2 Artificial Intelligence Combined with 5G Robotics
One thing that is certain is that this type of technology may enables robotics in a different way and ‘intelligent’ robots to operate within a broader environment. We can immediately consider the extended adverse use of this for defence purposes and monitoring a population. However, on the other hand within hospitals and health technology this may enable a different working environment or follow-up by technological inventions.
Running solutions related to the field of artificial intelligence is done externally already so will you truly notice a difference? If the video, text or other expression/input is transferred to a different place — then what is the difference for users. Using machine learning techniques in 5G network to improve the utility in the radio spectrum which is allocated. Currently it seems like Self-driving autonomous vehicles, Automating or managing public transportation, Time-critical industry automation, Remote healthcare. These are all areas where hardware and software will enable changes in operations. Of course process automations are already influencing working life with data from workers being monitored as well as performance measured. However, the robots that can operate in healthcare, transport and industry will likely be enhanced by these new technologies in Internet of Things (IOT).
3.3 Blockchain, 5G Robotics and AI
However much this title looks like a buzzword bingo there is something to be said for the security need required by these new type of solutions. Considering the question of who works in this IoT area we can argue that actors such as IOTA will see a great deal of growth going ahead.
“Our vision is to enable all connected devices through verification of truth and transactional settlements which incentivize devices to make available its properties and data in real time. This gives birth to entirely new general purpose applications and value chains.”
The Next Generation of Distributed Ledger Technology.
The IOTA protocol is a Distributed Ledger Technology developed by the IOTA Foundation. It is a next-generation…
www.iota.org
Therefore I think with the required security in these advancements company actors such as IOTA will be necessary to make it mesh or at least able to work.
3.4 A Looming Question — Sustaining 5G Robotics and AI
As mentioned in the introduction we must seriously question this advancement, we cannot continue running into ‘modernity’ and ‘advancement’ while putting both hands up to our ears. If we ignore the potentially adverse consequences of the aforementioned buzzword bingo we are in serious trouble. Energy and resource requirements by streaming AI to heavily encrypted 5G networks sounds absolutely ridiculous, is it even defensible to do so?
4. EU’s International Security Policy, Artificial Intelligence and 5G
4.1 Increased Connectivity and Emerging Risks
On the 9th of October 2019 Techcrunch wrote an article about a new joint risk assessment report into 5G technology. So far the European Union has resisted pressure to boycott Huawei as a 5G supplier on national security grounds. In March 2019 US suggested that EU should block Huawei from accessing its market. The report identifies “security challenges” more likely to appear in 5G networks.
“An increased exposure to attacks and more potential entry points for attackers: With 5G networks increasingly based on software, risks related to major security flaws, such as those deriving from poor software development processes within suppliers are gaining in importance.”
Certain equipment are becoming more sensitive such as base stations or management functions in networks. Reliance of mobile network operators on suppliers may lead to a higher number of ‘attack paths’. These can be exploited by threat actors. Non-EU States or State-backed are considered as the most serious ones and the most likely to target 5G networks. Risk profile of individual suppliers will become important. There is a risk of major dependencies on suppliers. Dependency can present a high degree of risk. Threat to availability and integrity as well as privacy and confidentiality.
According to TechCrunch the next step will be: “…the development, by December 31, of a toolbox of mitigating measures, agreed by the Network and Information Systems Cooperation Group, which will be aimed at addressing identified risks at national and Union level.”
“The implication is that the business of 5G security will need to get commensurately large to scale to meet the multi-dimensional security challenge that goes hand in glove with the next-gen tech. Just banning a single supplier isn’t going to cut it.”
4.2 Best British Tech Startups in 5G
KPMG hosts a competition for startups and within this I have seen advancements (I work with KPMG). I will list two of the startups mentioned in this competition:
Garrison Technology have pioneered an innovative security platform to counter the threat of persistent and targeted attacks, which exploit corporate internet connectivity. David Garfield, CEO at Garrison Technology, tells us more in his interview below. SenSat Technology uses simulated realities of major urban areas to extract actionable insights using machine learning techniques. This reduces the rollout cost for telecommunications companies for 5G and ultimately, the cost of 5G for consumers.
Security platforms is becoming important especially when the rollout of 5G handles sensitive data. Using AI to enhance these communications will be essential for robotics to be working well.
According to the security report: “…consideration should also be given to the development of the European industrial capacity in terms of software development, equipment manufacturing, laboratory testing, conformity evaluation, etc” Therefore we can expect more investment either into the industrial expansion in EU regarding 5G or a great deal of investment into 5G security over the coming years.
5. China and 5G in 2019
5.1 Racing Past Global Competition as a Leader in 5G Technology
China is said to be leading the charge on implementation of 5G, and I will explore an outline of the current situation towards the end of 2019. For this article I have read up on writing by Jeremy Goldkorn; a report by EY on China and 5G; recent articles in Bloomberg; and in Fortune. The official state-run press agency of the People’s Republic of China Xinhua reported earlier in 2019 about the current plans in China to expand the 5G network. It stated that the Ministry of Industry and Information Technology officially issued 5G commercial licenses to China Telecom, China Mobile, China Unicom and China Radio and Television.
Infographic from Xihuan news agency retrieved the 31st of October 2019
中国信息通信研究院《5G产业经济贡献》认为,预计20202025年,我国5G商用直接带动的经济总产出达10.6万亿元,5G将直接创造超过300万个就业岗位。
China Information and Communication Research Institute “5G Industry Economic Contribution” believes that from 2020 to 2025, China’s 5G commercial direct economic output will reach 10.6 trillion yuan, 5G will directly create more than 3 million jobs.
In this regard China also states that it welcomes foreign companies to work in China.
The ministries of Commerce and of Ecology and Environment released a short-term plan 2019–2020 to work on this issue together by promoting consumption of electronics and vehicles that depend on 5G. Commercially companies such as Huawei has begun moving fast towards developing their own technology to pull away from America with their ongoing trade war after being blacklisted by Trump.
A report by EY outlines that China is poised to win the ‘5G race’. They have made an apt infographic to illustrate this showing that China has a large share of the current 5G connections; revenue and a large degree of local connections.
Infographics by EY retrieved on the 31st of October.
Although trust in certain countries has dropped for China, there has still been an expressed interest to develop 5G together and there is an outline expenditure on 5G solutions.
The 30th of October the publication Bloomberg released an article titled China Rolls Out World’s Largest 5G Mobile Phone Network. The news related to China’s three state-owned wireless carriers debuted 5G mobile phone service. The article states:
“China Mobile Ltd., the country’s largest carrier, unveiled its network in 50 cities including Beijing, Shanghai and Shenzhen, with packages priced as low as 128 yuan ($18) a month”
This comes in a way against the backdrop of the rejection of US by Huawei and its 5G technologies. It was planned a year later, but is now rolled out earlier than planned. This means China has the largest commercial 5G network in the world.
Retrieved from Bloomberg on the 31st of October
Further Bloomberg writes that dominance in China is an important strategic move for Huawei:
“Dominance in the world’s largest market can blunt the effects of a U.S. campaign against other countries installing Huawei gear, which it accuses of posing a security threat. Despite the U.S. pressure, Huawei said in July that it had signed more than 60 commercial contracts to supply 5G networks around the world, including at least 28 in Europe.”
The consulting firm EY previously outlined a pathway to commercial launch, but it seems China is speeding ahead of these expectations.
Infographics by EY retrieved on the 31st of October
There is additionally a series of use cases to considered. I have mentioned some previously, however we can see that there is a larger amount of possibilities to be considered in personal and business use-cases.
In Fortune online Alan Segal, director of the Digital and Cyberspace Policy Program at the Council on Foreign Relations argued on the 31st of October 2019 that:
“A lot of it is also driven by the realization that (the U.S.) doesn’t have a manufacturer in this space, so, let’s try to slow the Chinese down,”
Combining this with one of the most ambitious policies for artificial intelligence in the world seems an interesting way to take technological leadership.
It seems at least that China has achieved 5G leadership, at least from a surface look at the issue.
In saying so we have to add with some moderation that knowledge about 5G and relating to these technologies and distributed. Size is not everything, and we may not know who is currently advancing 5G technologies beyond the current scope and large scale of the Chinese investment.
6. Fitbit + Google AI + 5G = ?
6.1 Google Buys Fitbit for $2.1 billion
An article in the Verge published the 1st of November 2019 announced the purchase of FitBit by Google:
Google has just announced that its buying wearable company Fitbit for $2.1 billion. In a blog post announcing the news, Google SVP of devices and services Rick Osterloh said that the Fitbit purchase is “an opportunity to invest even more in Wear OS as well as introduce Made by Google wearable devices into the market.”
This can be a complex investment, yet I have seen a few different arguments in terms of why this acquisition was made:
Google wants to invest in wearables and did not manage to develop solutions viable enough on its own. It is therefore buying FitBit to allow for this offering. Google is buying FitBit for proprietary technology. Google has bought FitBit for its data — not to increase ads, but to assist in its investment in health technology (which makes up large parts of the Alphabet portfolio).
There are likely more reasons, however there appears to be a strong awareness by the scare towards consumer in terms of protecting their health data. The statement directly from FitBit on the 1st of November said the following:
“Consumer trust is paramount to Fitbit. Strong privacy and security guidelines have been part of Fitbit’s DNA since day one, and this will not change. Fitbit will continue to put users in control of their data and will remain transparent about the data it collects and why. The company never sells personal information, and Fitbit health and wellness data will not be used for Google ads.”
Rick Osterloh, Senior Vice President, Devices & Services at Google. Said they would be bringing together the talent from FitBit in hardware together with software and artificial intelligence to build better wearables.
This happens simultaneously as Alphabet which owns Google is expanding its investments in Verily and Calico. These are two companies which are repeatedly mentioned in Alphabet Inc’s quarterly statements. Health data is a hot topic.
Although Google may not use FitBit data for ads it could very well use that aggregated data for drug development or insights for health in Verily and Calico. What stops Google/Alphabet from doing so?
I have fetched a description from the third quarterly (Q3) for 2019 page of Alphabet’s investor relations page. The following two sections are excerpts from Q3.
6.2 Calico
Calico is a life science company with a mission to harness advanced technologies to increase our understanding of the biology that controls lifespan.
In September 2014, AbbVie Inc. (AbbVie) and Calico entered into a research and development collaboration agreement intended to help both companies discover, develop, and bring to market new therapies for patients with age-related diseases, including neurodegeneration and cancer. In the second quarter of 2018, AbbVie and Calico amended the collaboration agreement resulting in an increase in total commitments. As of September 30, 2019, AbbVie has contributed $750 million to fund the collaboration pursuant to the agreement and is committed to an additional $500 million which will be paid by the fourth quarter of 2019. As of September 30, 2019, Calico has contributed $500 million and has committed up to an additional $750 million.
Calico has used its scientific expertise to establish a world-class research and development facility, with a focus on drug discovery and early drug development; and AbbVie provides scientific and clinical development support and its commercial expertise to bring new discoveries to market. Both companies share costs and profits for projects covered under this agreement equally. AbbVie’s contribution has been recorded as a liability on Calico’s financial statements, which is reduced and reflected as a reduction to research and development expense as eligible research and development costs are incurred by Calico.
As of September 30, 2019, we have contributed $480 million to Calico in exchange for Calico convertible preferred units and are committed to fund up to an additional $750 million on an as-needed basis and subject to certain conditions.
6.3 Verily
Verily is a life science company with a mission to make the world’s health data useful so that people enjoy healthier lives. In December 2018, Verily received $900 million in cash from a $1.0 billion investment round. The remaining $100 million was received in the first quarter of 2019. As of September 30, 2019, Verily has received an aggregate amount of $1.8 billion from sales of equity securities to external investors. These transactions were accounted for as equity transactions and no gain or loss was recognized.
6.4 What Do you Expect from $2.1 Billion?
When you buy a company for unimaginable sums of money (for most) it is not inconceivable that a return on investment is expected.
Then again according to Quartz: “Alphabet reported $32.6 billion in revenue from Google’s advertising business, a jump of about 16% over the same period last year.”
Still despite the large earnings that Alphabet has this is a rather sizeable bet. Fitbit which had seen its stock decline by ~80% since its IPO in 2015.
FitBit seems to have been declining for a few years prior to the purchase in terms of sales and revenue.
Google is buying its way into the wearables market and could increase drastically the expansion of FitBit as well as the possibilities of the platform.
Retrieved from IDC on the 2nd of November
6.5 The Bets Are On
Will this bet pay off? Does it matter?
This gives more power to understanding humans, which certainly has been part of Google’s success.
Amazon Alexa might speak to you, but with the developments in 5G getting a device on your body that can do so while consistently streaming the results may be possible to a higher degree.
Earlier this year it was predicted that 5G would boost the stock of FitBit:
“By enabling wearables to do so much more, and making them so much more indispensable, 5G will make wearables much more popular and valuable, boosting these companies’ financials.”
The advent of 5G will also enable wearables to more precisely monitor the physical, emotional, and mental well-being of individuals.
This investment by Google in other words is expected to pay off.
Another reason for the investment could be due to a realization that ‘search’ is changing. More people are asking Alexa and will be asking their own devices. The ‘battle’ or competition for devices is therefore possibly influencing Google’s current strategic investment.
Conclusion
5G networks are expanding and they hold potential, however we should not give in to the ‘race’-mentality. Although from a security perspective it is understandable, and it is a dilemma in terms of the environmental consequences. China is ‘leading the race’ in expanding the coverage into several cities, perhaps as a way to show that they are not intimidated by the ongoing trade war. 5G may enable smaller connected devices to work better in terms of their connections. Thus perhaps the move for Google buying heavily into wearables with the purchase of FitBit is a gamble that makes sense. However, the question of AI and 5G is a question as much, or even more, of international security as it is of personal security. There is no doubt in my mind now that these solutions are being rolled out they must be closely followed-up from within the companies, by government and independent researchers.

No comments:

Post a Comment