What does an AI think about Net Neutrality?

Originally published on my LinkedIn Newsletter, 9th Feb 2023. See here for comment thread

Two very important trends are occurring in tech I'm following at the moment, so I thought it might be fun to combine them:

• The emergence of #GenerativeAI, for answering questions, generating images and sounds, and potentially a whole lot more. OpenAI #ChatGPT is the current best-known, but there are dozens of others using language models, transformers & other techniques. Some people are suggesting it will redefine web search - and potentially an awful lot more than that. Some even see it as a pivotal shift in technology, society and "skilled" employment.
• The re-emergence of discussions around #NetNeutrality and associated regulation relating to technology platforms, telcos and networks, like the ridiculous (un)#fairshare & #InternetTrafficTax concept being pitched in Europe by lobbyists. In the UK, Ofcom recently concluded a consultation on whether changes to NN rules should be made (I sent in a reply myself - I'll discuss it another time).

So, I asked ChatGPT what it thought about NN, over a series of questions. I specifically focused on whether it helps or hinders innovation.

The transcript is below, but some thoughts from me first:

• The text is good. Almost alarmingly good. I found myself saying "good point" a few times. This is probably because it gives reasons in fairly long lists, not just "3 bulletpoints for a slide".
• It seems pretty even-handed, with "proponents say this, others say that"
• You can sense that its training base tends to give it "common" responses and frequently-cited arguments. That doesn't mean they're wrong, but it can lack nuance, or repeat some common *fallacies* as well.
• The long-list approach and comments is quite good at highlighting the need for multiple approaches & reconciling them. A single statistical approach isn't enough - there are also qualitative questions to ask, surveys to undertake and multiple angles to absorb. ChatGPT understands the issue is complex, even if it cannot resolve the complexity itself. Simple answers are wrong.
• I was particularly impressed with the response to my question on cross-border effects, and to separating NN vs. funding effects.
• It struggled with the last question, about Internet and non-Internet innovation. EDIT: I just realised that it ignored my typo, and understood that I meant "directly" when I wrote "direct" in my question
• I'd be very worried if I did bog-standard "content generation" for clickbait blogs, "bucketshop" market research, or corporate social posts. This is all perfectly-usable general blurb if you need slabs of text nobody is going to really delve into too deeply.

All that said, I don't think this is (yet) a gamechanger in terms of analysis and argument. It's good for collating a lot of "food for thought", but as others have pointed out, it can also be "confidently wrong", sometimes in amusing ways.

In a separate chat, I asked it for a contrarian's view of #5G use-cases. While it was right on some things ("Some argue that the vast majority of #IoT devices do not require the high speed and low latency that 5G provides and that the hype around IoT is overblown.") others were less convincing: ("Some argue that 5G may not provide the low latency and high reliability that is being promised, and that there are other technologies, such as wired connections, that are better suited for applications such as autonomous vehicles"). If there's one thing less relevant to self-driving cars than 5G, it's fixed broadband!

Anyway - enjoy the debate treanscribed in full, below.

(Note: the image for this article came from Stable Diffusion, when I asked for a picture showing "An advanced AI debating Net Neutrality")

Me: Does net neutrality help innovation?

ChatGPT: Net neutrality has been a topic of much debate, with proponents arguing that it promotes innovation and others arguing that it stifles it.

Proponents of net neutrality argue that it helps to level the playing field for new entrants and small businesses by preventing Internet Service Providers (ISPs) from discriminating against certain types of traffic or services. This, in turn, encourages innovation, as entrepreneurs and startups are free to bring new products and services to market without fear of being blocked or slowed down by ISPs.

On the other hand, opponents of net neutrality argue that it creates a regulatory burden that stifles innovation and investment in the broadband industry. They argue that without the ability to prioritize different types of traffic, ISPs will be unable to effectively manage their networks and may be forced to limit investment in new infrastructure and technologies.

Ultimately, the impact of net neutrality on innovation is a complex and highly debated issue, and there is no clear consensus on whether it helps or hurts innovation. However, proponents of net neutrality generally believe that it promotes a free and open internet that supports innovation and competition, while opponents believe that it imposes unnecessary regulation that stifles growth and investment.

Me: What research is necessary to reach a better conclusion?

ChatGPT: To reach a more definitive conclusion on the impact of net neutrality on innovation, several types of research would be necessary, including:

1. Economic studies: Economic studies could analyze the impact of net neutrality regulations on investment and innovation in the broadband industry, as well as the impact on consumer prices and access to broadband services.
2. Case studies: Case studies of countries or regions with and without net neutrality regulations could provide valuable insights into the real-world effects of these regulations on innovation and investment in the broadband industry.
3. Surveys and interviews: Surveys and interviews with entrepreneurs, startups, and industry experts could provide insights into the impact of net neutrality on the ability of new entrants to bring new products and services to market.
4. Technical studies: Technical studies could examine the impact of net neutrality regulations on network management and traffic prioritization, and assess the potential impact on network performance, reliability, and security.
5. Long-term data analysis: Long-term data analysis could track changes in investment, innovation, and consumer prices over time in regions with and without net neutrality regulations, providing valuable insights into the long-term impact of these regulations.

Overall, a comprehensive and multi-disciplinary approach to research, combining economic, case study, survey, technical, and data analysis methods, would be necessary to reach a more definitive conclusion on the impact of net neutrality on innovation.

Me: How would you measure innovation in this context?

ChatGPT: Measuring the impact of net neutrality on innovation is a complex task, as innovation can take many forms and can be difficult to quantify. However, several metrics could be used to measure innovation in this context, including:

1. Investment in research and development (R&D): This could be measured by tracking the amount of investment in R&D by broadband companies and startups, as well as the number of patents filed in relevant areas.
2. New product and service launches: This could be measured by tracking the number of new products and services introduced by broadband companies and startups, as well as the market adoption and success of these offerings.
3. Entrepreneurial activity: This could be measured by tracking the number of new startups founded in the broadband industry, as well as the success of these startups in attracting investment and bringing new products and services to market.
4. Consumer benefits: This could be measured by tracking changes in consumer prices, access to broadband services, and the availability of new and innovative online services and applications.
5. Network performance and reliability: This could be measured by tracking changes in network speed, latency, and uptime, as well as the ability of networks to handle new and innovative applications and services.

While these metrics are not exhaustive, they provide a good starting point for measuring the impact of net neutrality on innovation in the broadband industry. By tracking changes in these metrics over time in regions with and without net neutrality regulations, researchers could gain valuable insights into the impact of these regulations on innovation.

Me: A lot of those metrics are international. More regions with net neutrality might increase innovation globally, not locally. How do you account for that?

ChatGPT: You are correct, some of the metrics mentioned could be influenced by factors beyond the borders of a single region. To account for this, it would be important to consider both the local and global impact of net neutrality on innovation.

For example, to measure the local impact of net neutrality, researchers could compare the metrics mentioned in regions with and without net neutrality regulations, taking into account factors such as the local economy, cultural and regulatory environment, and the state of the broadband industry in each region. This would provide valuable insights into the specific impact of net neutrality on innovation within those regions.

To measure the global impact of net neutrality, researchers could compare these metrics on a global scale, taking into account factors such as the state of the global economy, the state of the broadband industry globally, and the regulatory environment in different regions. This would provide a broader perspective on the impact of net neutrality on innovation globally, and could help to identify any trends or patterns that may not be apparent from a regional analysis alone.

Overall, it is important to consider both the local and global impact of net neutrality on innovation to gain a complete and accurate understanding of the impact of these regulations on the broadband industry and innovation more broadly.

Me: How would we know if net neutrality had a larger effect on innovation than availability of financing for startups?

ChatGPT: Determining the relative impact of net neutrality versus the availability of financing on innovation can be challenging, as these factors are often interrelated and can have complex and interconnected effects. However, several approaches could be used to isolate the impact of each factor, including:

1. Comparative studies: Researchers could compare regions with and without net neutrality regulations, but with similar levels of financing available for startups. By controlling for the availability of financing, researchers could isolate the impact of net neutrality on innovation.
2. Historical analysis: Researchers could analyze data from periods of time when net neutrality regulations were in place or not in place, and compare this data with changes in the availability of financing for startups. This would help to identify any correlations between net neutrality and innovation, independent of changes in financing.
3. Causal inference methods: Researchers could use causal inference methods, such as regression analysis, to model the relationship between net neutrality, financing, and innovation, while controlling for other factors that could influence these relationships.
4. Survey and interview-based research: Researchers could survey and interview startups, entrepreneurs, and industry experts to gain insights into the relative importance of net neutrality and financing for innovation. This would provide valuable qualitative data to complement the quantitative data obtained from other research methods.

Overall, a multi-disciplinary and multi-method approach, incorporating comparative studies, historical analysis, causal inference methods, and survey and interview-based research, would be necessary to determine the relative impact of net neutrality and financing on innovation in a robust and comprehensive manner.

Me: How do you distinguish innovations that direct relate to Internet access, compared to those which involve non-Internet applications and technologies?

ChatGPT: Distinguishing innovations that directly relate to Internet access versus those that involve non-Internet applications and technologies can be challenging, as many innovations are influenced by both Internet access and non-Internet technologies. However, several approaches could be used to differentiate these types of innovations, including:

1. Technical classification: Innovations could be classified based on their underlying technologies, with innovations that rely primarily on Internet access being classified as Internet-related innovations and innovations that rely primarily on non-Internet technologies being classified as non-Internet-related innovations.
2. Market classification: Innovations could be classified based on the markets they serve, with innovations that serve primarily Internet-based markets, such as online commerce or online communication, being classified as Internet-related innovations and innovations that serve primarily non-Internet-based markets, such as manufacturing or agriculture, being classified as non-Internet-related innovations.
3. End-user classification: Innovations could be classified based on their end-users, with innovations that are primarily designed for use by individuals accessing the Internet being classified as Internet-related innovations and innovations that are primarily designed for use by individuals not accessing the Internet being classified as non-Internet-related innovations.

It is important to note that these classifications are not mutually exclusive, and that many innovations may have elements of both Internet-related and non-Internet-related technologies, markets, and end-users. Nevertheless, using these classifications can provide valuable insights into the extent to which innovations are influenced by Internet access, and can help to distinguish Internet-related innovations from non-Internet-related innovations.

 

Does the network need a "black box" as well as user data-retention?

What is the network equivalent of an aircraft's black-box? Is there an argument for governments pushing for more regulation on telco-side data-retention?

As far as I know, telcos are not under any obligation to maintain full logs of the state/operation of their network elements, either hardware or software – or make them available for authorities to inspect. As networks become more virtualised and complex, with NFV, orchestration, AI-led automation of network policies, slicing and so on, what happens if something goes seriously wrong? 

The industry is hoping that 5G and other networks will be used in safety-critical verticals, with "ultra-reliable" requirements, but that brings risks and responsibilities too.
That could mean authorities may need to do a diagnostic “post-mortem” if a network fails - or perhaps as a way to spot if the network is doing something it shouldn’t, such as discrimination in wholesale, or net neutrality violations.

Aviation has rigorous rules about flight data recorders (“black boxes”), and has an admirable record of learning lessons from catastrophe, and changing inspection and certification regimes, if needed. Air travel is a one-way ratchet, becoming ever-safer, because of this.

So, if a commercial 5G or FTTX network is being used for ultra-reliable uses (such as managing a power grid’s control, or a telemedicine app, or perhaps connected vehicles), is there a basis for countries having a “Network Accident Investigation Board” and better international cooperation? And would this not also imply a better way to store crucial background data is required? If a plane crashes, investigators can examine the physical wreckage, but this problem is much harder for software-controlled networks with no moving parts.

This is also an issue if a network gets compromised by hacking or a bug - who is responsible, how can it be fixed, and what prevents re-occurrence? And something similar applies for keeping records that may prove/disprove competition problems, eg did a virtualised network resource do something illegal, perhaps on a temporary basis? How could a complaint be investigated, or a prosecution brought?

The problems get multiplied massively if AI is involved, as any issues with underlying machine-learning algorithms are potentially a single point of failure, if that system is used widely (eg for coordinating 100’s or 1000’s of network-slices in an automated fashion).

Do regulators have the legal rights, obligation or ability to forensically analyse what’s gone wrong in such situations? Or the various cybersecurity agencies, or police forces?

One option might be to encrypt network configuration and operational logs, and keep them “in escrow” using blockchain to ensure anti-tamper properties, so that they could only be examined after a warrant or other legal instrument ordered decryption. There are likely numerous other technical approaches to consider as well.

In either case, as public networks become part of critical systems, this topic will only rise in importance. Policymakers should start thinking about it now - and the telecoms industry should face up to its responsibilities here, rather than push back without thinking. Do Boeing or Airbus complain about the need for data recorders?

Quick thoughts on 5G

I've been doing a lot of work - and events - on 5G recently. 

I've noticed a few recent shifts in perception and focus amongst vendors, regulators and operators. Some quick take-outs (a few more than appear on my similar LinkedIn post, as I'm not limited to 1300 characters!)

• 5G smartphones launch in 2019, but will be low-volume until 2020/21. Expect the first 5G iPhone towards the end of 2020
• Fixed-wireless use cases for 5G are high on the agenda in some markets (eg US, S Korea, Turkey, Germany), but seemingly almost absent in others.
• Commercial, large-scale, automated network slicing only becomes real from around 2023 onwards. A few "hand-carved" slices will be sooner, for example for internal use by MNOs' own business units, or perhaps public safety
• URLLC (ultra-reliable low latency) use-cases seem to have shifted from sci-fi fantasies around automated vehicles and surgical robots, to industrial IoT and factory automation... 
• ... but industrial use will often be controlled by industry itself, via one of several forms of private network, either using shared spectrum, private cores or private slices / enterprise MVNOs. MNOs' role may be minor
• Some claim that NB-IoT is the 5G version for "massive IoT", despite it being developed as a 4G variant. This is revisionist nonsense; if it was true then DT, VF and others would have been putting out PR 2+ years ago, claiming to be first to launch 5G
• 3.5GHz should be OK-ish outdoors but will struggle with outdoor-to-indoor coverage. mmWave will be worse. Beware of demos showing good indoor performance - ask about uplink from inside-out, or whether signals penetrate double-glazing, or at oblique angles to walls/windows. In any case, #WiFi will continue to dominate in the home.
• There will be some small-cells and neutral-host deployments for 3.5GHz (and similar bands) in enterprises and other large buildings, but this will take a long time to become widespread. 
• Existing in-building DAS systems will need some serious upgrades to support higher 5G frequency bands - most of today's top out at 2.6GHz and can't handle MIMO very well.
• Despite it not being an "official" 5G candidate band, 28GHz seems to be the most popular option, at least for test networks. This is partly because of chipset support, notably Qualcomm's X50. The European-proposed 26GHz hasn't seen much action yet
• Two of the largest 5G "verticals" associations, for Automotive (5GAA) and Industrial (5GACIA) seem to be heavily driven by German companies - and the German regulator looks like it's going to award 100MHz of spectrum to verticals directly (not 100% certain but getting clearer). In other countries apart from the US (CBRS) and China (Huawei's enterprise LTE), there doesn't seem to be as much action from large firms knocking on the regulator/governments doors.
• The 5G New Core is getting a lot of discussion and attention... but given that some of the existing NFV deployments have been slow, and the cost-savings somewhat illusory, I don't expect much near-term action on this.
• Some of the visions for 5G seem to lean heavily on automation and AI back-office for optimising radio, core, user-plane etc. Yet those are also still at an early stage - and few telcos have many skilled engineers -  so could act as a brake. There are also some emerging questions on security of network AI, and whether the algorithms might be single points of failure, especially when used for networks used for critical national infrastructure. 
• Connected-car companies are interested in 5G, but not as enthusiastic as some might imagine. One told me "it's a nice-to-have" - especially as vehicles will need to be able to work offline, and have prodigous on-board compute capabilities.
• I'm more positive about some of the discussion around Cloud RAN for 5G. In many ways, it's going to be necessary, given the complexity of NR. That said, there's some serious practical challenges about the radio, such as the size/weight/cost of the massive-MIMO antennas.
• There's lots of talk about network-slicing for 5G, but nobody has really thought about whether today's MNO wholesale departments are up to the task of selling "slice as a service". Speaking to some of today's MVNOs, it seems like they will have to do a lot of homework before they can become effective slicemongers.

That's a quick list of things off the top of my head. Plenty more observations and comments to come, or on my Twitter feed from various events I've attended.

If you'd like me to give an unvarnished presentation at an event, on "5G opportunities, realities & myths", please get in touch via:  information AT disruptive-analysis DOT com

And if you're interested in my last point, on 5G+MVNOs+Slicing+Wholesale, please look at my upcoming workshop doing a deep-dive on this (link)

MEC and network-edge computing is overhyped and underpowered

I keep hearing that Edge Computing is the next big thing - and specifically, in-network edge computing models such as MEC. (See here for a list of all the different types of "edge"). 

I hear it from network vendors, telcos, some consultants, blockchain-based startups and others. But, oddly, very rarely from developers of applications or devices.

My view is that it's important, but it's also being overhyped. Network-edge computing will only ever be a small slice of the overall cloud and computing domain. And because it's small, it will likely be an addition to (and integrated with) web-scale cloud platforms. We are very unlikely to see edge-first providers become "the next Amazon AWS, only distributed".

Why do I think it will be small? Because I've been looking at it through a different lens to most: power. It's a metric used by those at the top- and bottom ends of the computing industry, but only rarely by those in the middle, such as network owners. This means they're ignoring a couple of orders of magnitude.

(This is a long post. You might want to grab a coffee first....)


How many zeroes?

Cloud computing involves huge numbers. There are many metrics that you can use - numbers of servers, processors, standard-sized equipment racks, floorspace and so on. But the figure that gets used most among data-centre folk is probably power consumption in watts, or more commonly here kW, MW & GW. (Yes, it's a lower-case k for kilo). 

Power is useful, as it covers the needs not just of compute CPUs and GPUs, but also storage and networking elements in data centres. It's not perfect, but given that organising and analysing information is ultimately about energy it's a valid, top-level metric. [Hey, I've got a degree in physics, not engineering. Helloooo, thermodynamics & entropy!]

Roughly speaking, the world's big data centres have a total power consumption of about 100GW. A typical one might have a capacity of 30MW, but a number of the world's largest data centres already use over 100MW individually, and there are enormous plans for locations with 600MW or even 1GW (link). No, they're not all running at full power, all the time - but that's true of any computing platform.

This growth is partly driven by an increase in the number of servers and equipment racks needed (hence growing floor-space for these buildings). But it also reflects power consumption for each server, as chips get more powerful. Most equipment racks use 3-5kW of power, but some can go as high as 20kW if that power - and cooling - is available.

So, to power "the cloud" needs 100GW, a figure that is continuining to grow rapidly. We are also seeing a rise in smaller, regional data-centres in second- and third-tier cities. Companies and governments often have private data-centres as well. These vary quite a bit, but 1-5MW is a reasonable benchmark.


How many decimal places?

At the other end of the computing power spectrum, are devices, and the components inside them. Especially for battery-powered devices, managing the power-budget down to watts or milliwatts is critical. This is the "device edge".

• Sensors might use less than 10mW when idle & 100mW when actively processing data
• A Raspberry Pi might use 0.5W
• A smartphone processor might use 1-3W
• An IoT gateway (controlling various local devices) might be 5-10W
• A laptop might draw 50W
• A decent crypto mining rig might use 1kW

New innovations are pushing the boundaries. Some researchers are working on sub-milliwatt vision processors (link). ARM has designs able to run machine-learning algorithms on very low-powered devices.

But perhaps the most interesting "device edge" is the future top-end Nvidia Pegasus board, aimed at self-driving vehicles. It is a 500W supercomputer. That might sound a lot, but it's still less than 1% of the engine power on most cars. A top-end Tesla P100D puts over 500kW to the wheels in "ludicrous mode", or 1000x that figure. Cars' aircon might use 2kW, to give context.

Of course, all of these device-edge computing platforms are numerous. There are billions of phones, and hundreds of millions of vehicles and PCs. Potentially, we'll get 10s of billions of sensors. Most aren't coordinated, though. 


And in the middle?

So we have milliwatts at one end of distributed computing, and gigawatts at the other, from device to cloud.

So what about the middle, where the network lives?

There are many companies talking about MEC (multi-access edge computing) and fog-computing products, with servers designed to run at cellular base stations, network aggregation points, and also in fixed-network nodes and elsewhere. 

Some are "micro-data-centres" capable of holding a few racks of servers near the largest cell towers. The very largest might be 50kW shipping-container sized units, but those will be pretty rare and will obviously need a dedicated power supply.

It's worth noting here that a typical macro-cell tower might have a power supply of 1-2kW. So if we consider that maybe 10% could be dedicated to a compute platform rather than the radio (a generous assumption), we get 100-200W, in theory. Or in other words, a cell tower edge-node will be less than half as powerful as a single car's computer.

Others are smaller server units, intended to hook into cellular small-cells, home gateways, cable street-side cabinets or enterprise "white boxes". For these, 10-30W is more reasonable.

Imagine the year 2023

Let's think 5 years ahead. By then, there could probably be 150GW of large-scale data centres, plus a decent number of midsize regional data-centres, plus private enterprise facilities.

And we could have 10 billion phones, PCs, tablets & other small end-points contributing to a distributed edge, although obviously they will spend a lot of time in idle-mode. We might also have 10 million almost-autonomous vehicles, with a lot of compute, even if they're not fully self-driving. 

Now, imagine we have a very-bullish 10 million "deep" network-compute nodes, at cell sites large and small, built into WiFi APs or controllers, and perhaps in cable/fixed streetside cabinets. They will likely have power ratings between 10W and 300W, although the largest will be numerically few in number. Choose 100W on average, for a simpler calculation. (Frankly, this is a generous forecast, but let's run with it for now).

And let's add in 20,000 container-sized 50kW units, or repurposed central-offices-as-datacentres, as well. (Also generous)

In other words, we might end up with:

150GW large data centres
50GW regional and corporate data centres
20,000x 50kW = 1GW big/aggregation-point "network-edge"
10m x 100W = 1GW "deep" network-edge nodes
1bn x 50W = 50GW of PCs
10bn x 1W = 10GW "small" device edge compute nodes
10m x 500W = 5GW of in-vehicle compute nodes
10bn x 100mW = 1GW of sensors & low-end devices

Now admittedly this is a very crude analysis. And a lot of devices will be running idle most of the time, and may need to offload functions to save battery power. Laptops are often switched off entirely. But equally, network-edge computers won't be running at 100%, 24x7 either.


The 1% edge

So at a rough, order-of-magnitude level, we can see that the total realistic "network edge", with optimistic assumptions, will account for less than 1% of total aggregate compute capability. And with more pessimistic assumptions, it might easily be just 0.1%. 

Any more will simply not be possible to power, unless there are large-scale upgrades to the electricity supply to network infrastructure - installed at the same time as backhaul upgrades for 5G, or deployment of FTTH. (And unlike copper, fibre can't even power small devices on its own). And haven't seen announcements of any telcos building hydroelectric power stations anywhere.

Decentralised, blockchain-based edge "fogs" are unlikely to really solve this problem either, even if they also use decentralised, blockchain-based power supply and management.

Now it could be argued that this 0.1-1% of computing workloads will be of such pivotal importance, that they will bring everything else into their orbit and indirect control. Could the "edge" really be the new frontier? 

I think not.

In reality, the reverse is more likely. Either device-based applications will selectively offload certain workloads to the network, or the webscale clouds will distribute certain functions. Yes, there will be some counter-examples, where the network-edge is the control point for certain verticals or applications - I think some security functions make sense, for instance, as well as an evolution of today's CDNs. But will IoT management, or AI, be concentrated in these edge nodes? It seems improbable.


Conclusion & TL:DR

In-network edge-computing architectures, such as MEC, will become more important. There are various interesting use-cases. But despite that, they will struggle to live up to the hype. 

There will be almost no applications that run *only* in the network-edge - it’ll be used just for specific workloads or microservices, as a subset of a broader multi-tier application. The main compute heavy-lifting will be done on-device, or on-cloud. As such, collaboration between edge-compute providers and industry/webscale cloud will be needed, as the network-edge will only be a component in a bigger solution, and will only very rarely be the most important component. 

One thing is definite: mobile operators won’t become distributed quasi-Amazons, running image-processing for all nearby cars or industry 4.0 robots in their networks, linked via 5G. 

Yes, MEC nodes could host Amazon Greengrass or other functions on a wholesale basis, but few developers will want to write directly to telcos' distributed-cloud APIs on a standalone basis, with or without network-slicing or 5G QoS mechanisms.

Indeed, this landscape of compute resource may throw up some unintended consequences. Ironically, it seems more likely that a future car's hefty computer, and abundant local power, could be used to offload tasks from the network, rather than vice versa.

Comments and feedback are very welcome. I'm aware I've made many assumptions here, and will doubtless generate various comments and detailed responses, either on my blog or LinkedIn posts. I haven't seen an "end to end" analysis of compute power before - if there's any tweaks to my back-of-envelope calculations, I'd welcome suggestions. If you'd like to contact me about projects or speaking engagements, I can be reached via information at disruptive-analysis dot com.

Telco use-cases for AI: A simple categorisation model

The coming years will see the application of AI technology across all sectors of the economy and life. The telecoms industry is no different. Although I’ve been commenting on telco-sector AI in the context of “TelcoFuturism” for some time (link), and co-ran a workshop on it in May 2017 (link), the last few months have seen a notable upswing in interest. I’d say that the public use-cases now seem to be significantly in advance of those for blockchain, in terms of potentially-transformational technologies.

That said, it can still be hard for many executives to grasp exactly what is likely to change, and when, for AI/telecoms combinations. This is highlighted by the surge in AI-related panels, presentations and even complete streams at industry conferences – although sometimes I see more interest from generalist AI people about the telecoms vertical, versus telecom specialists looking at what’s new. 

Both sides of the equation have large volumes of obscure acronyms, multi-layered technology stacks, and complex volume chains – which can mean that mutual understanding is often confined to narrow niches. AI covers machine- and deep-learning, language processing, machine-vision and much more. Telecoms includes vast realms of internal systems and processes that are unknown to most who are not insiders – domains like core networks, OSS/BSS, network optimisation, toll fraud and service-assurance are alien to those not steeped in the industry.

One of the ways I’ve been using to “set the scene” for describing AI/telecoms intersections is to simplify and categorise the use-case areas. I count three, possibly four, large “buckets” into which a variety of telecom AI impacts will fit. These buckets are not based on either specific AI or telecoms technology slices, but more on understandable business functions and roles:

• Dealing with customers
• Managing operations
• Creating new services
• (External risks)

 Within each of these areas, there are many, many sub-sectors – and also some overlap.

“Dealing with customers” can include everything from voice/text chatbots for customer-service, through to predictions of which customers are least-happy and may “churn” to competitors. Where telcos have retail outlets, it could incorporate various in-store technologies, or it could be about smarter web-consoles for B2B customers running complex managed services.

“Managing operations” is even more diverse – it could be fault prediction for network elements, optimising the 100s of configuration variables for radio networks, spotting fraudulent traffic to international premium-rate numbers, allocating engineering resources more productively, or protecting against hackers and malware. There are hundreds of possible uses here, which mostly overlay on top of existing operational/business support systems (OSS/BSS) See also my recent post (link)

“New services” also spans a range of areas, but broadly splits between AI-enabled and AI-enabling services. An AI-enabled service could be a local-language voice assistant added to a cable operator’s set-top box or remote control. Or it could be the provision of integrated “smart city” solutions including video-cameras and security analytics. AI-enablement could include offering “edge” servers for hosting local processing, milliseconds transport-time away from a device, or it could be the provision of anonymised bulk data for others to apply algorithms to. Telco opportunities with IoT+AI include both enablement and enabled services, in numerous manifestations.

The “risks” category includes a diffuse set of possibilities by which AI might harm the telecom industry, or dampen demand for services. Smarter devices (eg autonomous vehicles) will be able to host their own offline image processing & route-planning locally, rather than needing realtime connectivity at 5G speeds/latencies. Another threat could be customers’ smart assistants renegotiating price-plans on their behalf – after crowdsourcing millions of conversations to deduce how best to game the retention staff’s scripts and objections. (Of course in the latter example, the customer-retention team could themselves be bots). Numerous types of automated “least-cost X” and arbitrage engines are likely to emerge. Various security risks are also probable here too.

Clearly, using just these four "buckets" misses much of the fine-grained detail. But I find it helpful as a starting point, as most top-level industry issues apply differently to each. 

Consider input data, for example – for both customer management and operations, telcos have abundant historical records and ongoing data collection that may generate terabytes per day. But for the former, privacy considerations often come to the fore in terms of regulation and risk, while this is far less of a concern for internal operational data, for example on how the network is running. For new services, almost by definition the focus is on collecting/processing/transporting new data, rather than deriving conclusions from existing sets. 

This four-way framework is also useful for thinking about different types of ROI model - split broadly between impacting existing revenues, existing costs, new revenues and potential changes to underlying assumptions. 

I'll be covering these topics in more depth in various upcoming presentations and reports, as well as looking at other areas of telco-linked innovation such as blockchain, 5G and enterprise verticals. Please get in touch if you would like more detail, or are interested in internal workshops, external support through events or white papers, or are seeking ongoing strategic advisory support.