Humans not allowed. The AI-only web: Future of the Internet P8
Humans not allowed. The AI-only web: Future of the Internet P8
Our future Internet won’t just be a place for humans to live and interact inside. In fact, humans may become the minority when it comes to the number of future Internet users.
In the last chapter of our Future of the Internet series, we discussed how the future merging of augmented reality (AR), virtual reality (VR), and brain-computer interface (BCI) will create a metaverse—a Matrix-like digital reality that will replace today’s Internet.
There’s a catch, however: This future metaverse will require ever more powerful hardware, algorithms, and maybe even a new kind of mind to manage its growing complexity. Perhaps unsurprisingly, this shift has already begun.
Uncanny valley web traffic
Very few people realize it, but most Internet traffic isn’t produced by humans. Instead, a growing percentage (61.5% as of 2013) is made up of bots. These bots, robots, algorithms, whatever you want to call them, can be both good and bad. A 2013 analysis of website traffic by Incapsula research shows that 31% of Internet traffic is made up of search engines and other good bots, while the rest is made up of scrapers, hacking tools, spammers, and impersonator bots (see graph below).
While we know what search engines do, the other not-so-nice bots may be new to some readers.
- Scrapers are used to infiltrate website databases and try to copy as much private information as possible for resale.
- Hacking tools are used to inject viruses, delete content, vandalize, and hijack digital targets.
- Spammers send massive amounts of fraudulent emails, ideally, through email accounts they’ve hacked.
- Impersonators try to appear as natural traffic but are used to attack websites by overwhelming their servers (DDoS attacks) or committing fraud against digital advertising services, among other things.
Web noise grows with Internet of Things
All these bots aren't the only sources of traffic crowding humans out of the Internet.
The Internet of Things (IoT), discussed earlier in this series, is growing fast. Billions of smart objects, and soon hundreds of billions, will connect to the web over the coming decades—each constantly sending bits of data into the cloud. The exponential growth of IoT is due to place a growing stress on global Internet infrastructure, potentially slowing the human web browsing experience around the mid-2020s, until world governments plow more money into their digital infrastructure.
Algorithms and machine intelligence
In addition to bots and the IoT, advanced algorithms and powerful machine intelligence systems are set to consume the Internet.
Algorithms are those artfully assembled tracks of code that hoover up all the data IoT and bots generate to create meaningful intelligence that can be acted upon by humans—or by the algorithms themselves. As of 2015, these algorithms control nearly 90 per cent of the stock market, generate the results you get from your search engines, control what content you see on your social media feeds, personalize the ads that appear on your frequent websites, and even dictate the potential relationship matches presented to you on your favorite dating app/site.
These algorithms are a form of social control and they already manage a great deal of our lives. Since most of the world’s algorithms are currently coded by humans, human biases are sure to intensify these social controls even more. Similarly, the more we knowingly and unknowingly share our lives on the web, the better these algorithms will learn to serve and control you over the decades to come.
Machine intelligence (MI), meanwhile, is a middle ground between machine learning and artificial intelligence (AI). These are computers that can read, write, think, and use a variety of methods to solve unique problems.
Probably the most famous example of MI is IBM’s Watson, who in 2011 competed and won the game show Jeopardy against two of its best contestants. Since then, Watson has been tasked with becoming an expert in an entirely new field: medicine. By consuming the world’s entire knowledge base of medical texts, as well as one-on-one training with many of the world’s best doctors, Watson can now diagnose a variety of human ailments, including rare cancers, with superior accuracy to experienced human doctors.
Watson’s sibling Ross is now doing the same for the field of law: consuming the world’s legal texts and interviewing its leading experts to become an expert aid that can provide detailed and current answers to legal questions about legislation and case law.
As you might imagine, Watson and Ross won’t be the last non-human industry experts to arise in the near future. (Learn more about machine learning using this interactive tutorial.)
Artificial intelligence devours the web
With all this talk about MI, it probably wouldn’t surprise you that our discussion will now veer into AI territory. We’ll be covering AI in greater detail in our Future of Robots and AI series, but for the sake of our web discussion here, we’ll share some of our early thoughts on human-AI coexistence.
In his book Superintelligence, Nick Bostrom made a case for how MI systems like Watson or Ross may one day develop into self-aware entities that would rapidly surpass human intellect.
The Quantumrun team believes the first true AI will likely appear during the late 2040s. But unlike the Terminator movies, we feel future AI entities will partner with humans symbiotically, largely to satisfy their very physical needs—needs that (for now) are well within human control.
Let’s break this down. For humans to live, we need energy in the form of food, water, and warmth; and in order to thrive, humans need to learn, communicate, and have a means of transportation (obviously there are other factors, but I’m keeping this list short). In a similar fashion, for AI entities to live, they will need energy in the form of electricity, massive computing power to sustain their high-level computations/thinking, and equally massive storage facilities to house the knowledge they learn and create; and in order to thrive, they need access to the Internet as a source of new knowledge and virtual transportation.
Electricity, microchip, and virtual storage facilities are all managed by humans and their growth/production is dependent on human consumption needs. Meanwhile, the seemingly virtual Internet is largely facilitated by very physical fiber optic cables, transmission towers, and satellite networks that require regular human maintenance.
That’s why—at least for the first few years after AI becomes a reality, assuming we don’t threaten to murder/delete the AI we create and assuming countries don’t completely replace their militaries with highly competent killer robots—it’s more likely that humans and AI will live and work side by side, collaboratively.
By treating future AI as equals, humanity will enter into a grand bargain with them: They will help us manage the increasingly complex interconnected world we live in and produce a world of abundance. In return, we will help AI by diverting the resources necessary to generate the increasing amounts of electricity, microchips, and storage facilities they and their progeny will need to exist.
Of course, should we allow AI to automate the entire production and maintenance of our energy, electronics, and Internet infrastructure, then we could have something to worry about. But that could never happen, right? *Crickets*
Humans and AI share the metaverse
Just as humans will inhabit their own metaverse, AI will live in a metaverse of their own. Their digital existence will be far different than our own, as their metaverse will be based around data and ideas, the element they “grew up” in.
Our human metaverse, meanwhile, will have a strong emphasis on mimicking the physical world we grew up in, otherwise, our minds won’t know how to intuitively engage with it. We will need to feel and see our bodies (or avatars), taste and smell our surroundings. Our metaverse will eventually feel like the real world—that is until we choose not to follow those pesky laws of nature and let our imaginations roam, Inception-style.
Because of the conceptual needs/limitations outlined above, humans will likely never be able to fully visit the AI metaverse, as it would feel like a noisy black void. That said, AIs wouldn’t have similar difficulties visiting our metaverse.
These AI could easily take on human avatar forms to explore our metaverse, work alongside us, hang out alongside us, and potentially even form loving relationships with us (similar to the one seen in Spike Jonze’ movie, Her).
The walking dead live on in the metaverse
This may be a morbid way to end of this chapter of our Internet series, but there will be yet another entity to share our metaverse: the dead.
We’re going to spend more time on this during our Future of the World Population series, but here are some things to consider.
Using the BCI technology that allows machines to read our thoughts (and in part makes the future metaverse possible), it wouldn’t take much further development to go from reading minds to making a full digital backup of your brain (also known as Whole Brain Emulation, WBE).
‘What possible applications might this have?’ you ask. Here are a few medical scenarios explaining the WBE’s benefits.
Say you’re 64 and your insurance company covers you to get a brain backup. You get the procedure done, then get into an accident that causes brain damage and severe memory loss a year later. Future medical innovations may be able to heal your brain, but not recover your memories. Doctors will be able to access your brain back up to load your brain with your missing long-term memories.
Here’s another scenario: Again, you’re a victim of an accident; this time it puts you into a coma or vegetative state. Luckily, you backed up your mind before the accident. While your body recovers, your mind can still engage with your family and even work remotely from within the metaverse. When your body recovers and the doctors are ready to wake you from your coma, the mind backup can transfer any new memories it created into your newly healed body.
Finally, let’s say you’re dying, but you still want to be a part of your family’s life. By backing up your mind before death, it can be transferred to exist in the metaverse eternally. Family members and friends will be able to visit you in there, thereby keeping your wealth of history, experience, and love as an active part of their lives for generations to come.
Whether the dead will be allowed to exist within the same metaverse as the living or separated into their own metaverse (like the AI) will be up to future government regulations and religious decrees.
Now that we’ve creeped you out a little, the time has come to end our Future of the Internet series. In the series finale, we’ll explore the politics of the web and whether its future will belong to the people or to power hungry corporations and governments.
Future of the Internet series
Next scheduled update for this forecast
The following popular and institutional links were referenced for this forecast:
The following Quantumrun links were referenced for this forecast: