AI: Watch TikTok and others

...for the Net Good

AI In My View, 7/23/23

This Sunday, let’s focus on two consecutive weekend op-ed pieces by geopolitical commentator Fareed Zakaria, which revolve around TikTok. The globally recognized entertainment app, TikTok, has garnered global attention with its AI-driven “For You” algorithm, captivating hundreds of millions across demographics. Now emulated by tech companies large and small.

Fareed’s first op-ed on April 7, 2023, titled “Why banning TikTok won’t do any good”. He followed a week later on April 14 with “TikTok is dangerously addictive. We should regulate it now.”

Both are worth reading, but let me summarize why I’m highlighting them together as bookends now. 

The first piece makes the case that the current unusual bipartisan zeal to bash and potentially ban TikTok for its parentage may not be productive.  It’s owned by Chinese mega tech Bytedance, and there’s concerns around potential Chinese government influence and control.  Specifically, I agree with how he ends the piece:

“For years, we mistakenly believed that as China opened up economically it would become more like us. But look at where we are today.”

“We’ve embarked on central economic planning with massive subsidies to industry, and now we’re proposing draconian restrictions on the free flow of information. It seems like we are slowly becoming more like them.”

I view the case against China as politically overwrought. And likely distracting given that our mega tech and telecommunications companies, are doing many of the same things in a deeper way (sans of course the Chinese Communist party influence). 

I also believe that the general US China tensions are politically over wound. And it’s likely in our best interests to be long term focused in figuring out how we can achieve our global growth objectives with China, than without. Thread the needle as it were. 

I know it’s not the popular consensus, but I am convinced that the Math ultimately doesn’t work without China for us. I’ll have more on this subject in future posts. 

For now I want to focus on Fareed’s second piece above. Specifically, his view:

“Last week, I argued against banning TikTok. In talking to people about the platform, I came to see that the real concern most had was not about TikTok’s Chinese ownership, but rather just how scarily addictive it — and much of social media — is. That’s true and deeply worrying, and we should do something about it — and soon.”

The emphasis beyond TikTok, and on “much of social media”, is appropriate. 

TikTok is just the tip of the AI/ML iceberg,  applied to consumer apps at scale. As I’ve outlined in earlier posts, I view TikTok as the best example at scale to date, of the power of machine learning and AI algorithms (not to mention cleverly licensed universally loved music), applied to consumer-facing applications to date. 

And we’re in early days of AI. Before being turbocharged soon by LLM AI versions much more powerful. And of course it’s beyond TikTok, because companies large and small are rushing down the same path. 

Meta, OpenAI/Microsoft, Google, eventually Apple, and many ‘AI Native’ startups are also leading the charge. My recent posts have gone into these details. 

Fareed’s core points on TikTok apply to AI being developed by many other companies at large. TikTok is just one of the first to wield it effectively at scale, eventually for net good. 

Taking a step back, most technologies are double edge swords. There are perennial ways to use them for both bad and good. This has been true since Prometheus and Fire.

The key always lies in how society manages and navigates the complexities introduced by new technology, using it responsibly for the greater good. It takes time. And a LOT of trial and error. 

And it all needs to do be done as individuals, as companies, as governments. Whether it’s for personal pleasure or profit, corporate pursuit of invisible hand driven growth , or government use of the technology for politics, national security, sovereign, and societal good.

AI is not intrinsically a malevolent thing despite our worst fictional fables. It’s but a tool, and the only way it ultimately does any good, is if it’s used by the three constituents above for net good over time. And then they all continue that net good, across national boundaries. 

That’s how we figured out how to use air travel to help connect the world as the safest and fastest technology for net good. And shipping containers in the 1950s, and ships to enable globalization and do net good across the world for countries both rich and poor. 

And in the process lift more people out of abject poverty than any time in history. Looking at you of course China, for getting that right. 

Insanely hard working and competing Chinese people drove that at scale. Others preceded at a lesser scale, and many are fast following. 

And of course similar examples abound in healthcare, energy, manufacturing and almost every industry that has wielded technology for net global good. 

AI has the promise to augment all of us individually, organizationally, and globally. It only happens with a lean-in attitude into that opportunity, and being clear-eyed on the fears and risks. And it certainly is not going to happen in a straight line. 

Technologies like radio, TV (broadcast and cable), PCs, the Internet, and now AI, oscillate human existence between Aldous Huxley’s ‘Brave New World’, and George Orwell's '1984’. This was brilliantly summarized by Neil Postman’s ‘Amusing ourself to Death’. Technology driven blissful entertainment, progress, and societal control, going hand in hand. 

It’s notable that the focus of all three of those seminal works is intrinsically about ‘Attention’. Society’s attention and government control of attention, again going hand in hand. 

So it’s serendipitously appropriate that this phase of the AI Tech wave was kicked off by an AI paper on Attention. Am referring of course, to the seminal 1917 Google AI paper ‘Attention is all you need’. It taught us how to transform computers from deterministic code to probabilistic code processors at unimaginable scale.

We use these AI ‘Transformer’ algorithms, to calculate massively around every speck of human generated digital Data, using Attention driven weights, tokens, contexts, and reinforcement learning loops. And the technologies are barely getting started. 

But we ultimately have the power to figure out how to use AI for the right things (aka ‘the good’), after trying everything else (lot of ‘the bad’). Net Good. Just like America. Stay tuned. 

==

Finally, on a separate note, the fact that this post starts out inspired by two Fareed Zakaria weekend pieces, is not a coincidence. 

I’ve been a long time fan of Fareed’s ‘Here’s My Take’ segments on his GPS CNN TV show, and his Washington Post op-eds.

So this post titled ‘AI In My View’, is in the same vein. An opportunity to talk about AI from the broader vantage point in terms of societal net good. 

Above the daily AI din and details during the week, on who’s up and who’s down. Am planning on doing these on Sundays. Hopefully these will be entertaining and useful. Thanks for reading this far.