TikTok's new CEO says company will reveal how its algorithms work

>>When they finally release the algorithm how will we know that's the whole story?

Bingo. Unless you have a NSA squad checking every code change they make. Don't use them and be done.

The truth may be some combination of the two - it's certainly arguable that more transparent social network algorithms could be in everyone's interests; it's also true that TikTok would like to find ways to compete with Facebook, and increasing access to their source code may be part of their approach.

Attempting to take a global view, the ability for users themselves to scrutinize source code could be particularly important if & when subsets of them encounter bugs or get stuck in software that increasingly mediates conversations and real-world interactions.

In any case, as you say, source code is only part of the equation when we're determining how trustworthy the companies we deal with are.

As it happens, Facebook mentioned some of their contributions to the open source community as part of their testimony[1] to the Antitrust committee hearing today:

"We actively contribute to the open-source community. ... I believe sharing our intellectual property this way helps the entire ecosystem move forward and develop new products."

The TikTok Transparency Center[2] announced by the CEO does not yet include access to source code, as far as I can tell, although it does mention it.

[1] - https://docs.house.gov/meetings/JU/JU05/20200729/110883/HHRG...

[2] - https://www.tiktok.com/transparency

> When they finally release the algorithm how will we know that's the whole story?

We already know the algorithm won't be the whole story; it already isn't.

https://news.ycombinator.com/item?id=22610829

There's also a section on censorship in the middle of that FA. My younger brother who does lots of TikTok says that he doesn't see anything about China, good or bad; but he has seen plenty of the stereotypical Chinese anti-USA stuff (coverage on riots, unrest, etc.). But of course that could just be because he doesn't care about China and doesn't view Chinese stuff on TikTok. Anyway, we can't tell anything unless the moderation process and decisions are also made public.

Clearly though neither Apple nor Google have thought TikTok problematic enough to not delist it from their app stores.

I don't think any app should be banned unless they indulge in criminality/illegality which needs to be proven in law.

> When they finally release the algorithm how will we know that's the whole story?

Censorship by moderators are beyond the algorithms. Their algorithms and moderators censored Tiananmen Square, Tibetan independence, Falun Gong [0], Uighur sympathizer [1], Hong Kong protests [2], and (as another commenter points out) the ugly, poor, or disabled [3].

Their moderation guidelines banned contents “endangering national security” or deemed “uglification or distortion of local or other countries’ history” [4].

[0] https://www.theguardian.com/technology/2019/sep/25/revealed-...

[1] https://www.cnbc.com/2019/11/26/tiktok-says-it-doesnt-censor...

[2] https://www.washingtonpost.com/technology/2019/09/15/tiktoks...

[3] https://theintercept.com/2020/03/16/tiktok-app-moderators-us...

[4] https://www.socialmediatoday.com/news/tiktok-once-again-come...

The american government isn't using the data to openly oppress entire groups of people though.

This is clearly over more than TikTok, it's just the most convenient target.

I can see both sides. The US doesn't want Chinese tech because China won't allow US tech. That's really what it boils down to. And I can't really blame the US for it because China has locked out US tech companies for decades.

What role does Singapore play in this? Genuinely curious and I do not have nor recommend the app to anyone.

>Its the data collection

Exactly, it's not about relevancy of content, is what it's being done with the data.

I mean unless they are trying to pull the hook of "no you can see we don't give any benefit to pro-ccp/china/anti-western content, so this can't be a ccp propaganda machine!"

We know there's propaganda there, it's on every social media.

The problem is what's being harvested and shared with the CCP.

There may be concerns that TikTok's algorithms manipulate content to push certain trends and control public opinion so hypothetically there could be value in releasing the algorithms. They could also just as easily lie and leave out important sections. At this point there is no point except for empty posturing because no one will trust anything they release.

From the article: "He (Kevin Mayer) says TikTok will also reveal its data flows to regulators, and is calling on its rivals to do the same."

I could see a malicious recommendation algorithm being a large problem, but even if that did exist there is no way they would show that. By malicious, I mean one that identifies content and then recommends it in a way that would achieve a goal people would not support TikTok openly supporting. The simplest is if it automatically ranked content as pro or anti China and used that to influence the recommendations, so for two similar pieces of content with regards to what the user has indicated they want to see, the more pro-china one would be shown. A small bias would be hard to detect and wouldn't achieve much for any given person, but given tens of millions of users over a few years and it would cause a noticeable shift. Other options include slightly promoting content that would cause a social/moral decay in targeted countries. Examples of that would include a slight favoritism to content that promotes violence as a means for conflict resolution or a disregard of a users age when deciding to show or allow for upload more purient content. Once again, almost no impact on a single user, but potentially enough to cause a social shift.

That's not to say any of this is happening. I was just coming up with hypothetical ways an intelligent enough recommendation/moderation algorithm could be harmful even without the whole data collection aspect.

Like you I doubt they will release the entire thing. I think this is more of a marketing stunt than an actual drive for more trust. Less tech savvy users will see this as "transparency" and that will drive more trust.

The thing is, once tiktok has captured enough of the userbase, there may be no necessity of protecting the secret sauce. If coke release their formula, coke would still remain dominant even if exact duplicates were pushed by startup cola factories.

Literally in the first paragraph

> He says TikTok will also reveal its data flows to regulators, and is calling on its rivals to do the same

Can someone please share a source on how it is more of a spyware than facebook, instagram, etc.

Edit: grammar

> app is effectively spyware

What is your source for saying that? And what is your definition of spyware?

I don't use TikTok, Facebook, nor Instagram. But I don't get how TikTok is different than the other two? I don't call Facebook a spyware but I don't use it for the very similar reasons that I don't use TikTok and Instagram.

Are we worry about fab or google because of their ranking system? What tik tok can do is they can change the flip to send info over to their choices. That is the same as 5G. It is not what it does today but it can do on the next software or firmware upgrade will. The regime is so different it is a war our there already. Just you cannot see gunshot in software. But you can see casualties at least like HONG kong.

CEOs are out in force today.

I see another problem, and that is the one where they dump a model on the internet somewhere and whenever something goes wrong shrug their shoulders and say "well, that is what the model gave us…we don't know why; you can see for yourself how we didn't bias anything".

And if they are today, who knows what code changes are going in the pull request for the next day

Couldn't the userbase be more to blame than Twitter?

Pew Research says that "Twitter users [are] more likely to be Democrats" [1]. Also, 80% of all Tweets mentioning national politics were apparently made by users who "strongly disapprove" of Trump [2]. That provides a potential explanation for why comments and trending tweets may "favor Democrat issues" -- supposing that even is the case.

I can't seem to find any reputable articles or evidence regarding the "trending tweets" political breakdown and "campaign contributions by the staff of Twitter" -- could we be pointed to some sources?

[1] https://www.pewresearch.org/internet/2019/04/24/sizing-up-tw... [2] https://www.pewresearch.org/politics/2019/10/23/national-pol...

I doubt people believe. It just shows the power imbalance, they lie to the face and no one can do anything about it.

The West isn't uniform on this at all. Eastern Europe and up to Greece and Italy have taken a fairly conciliatory stance on China, in particular the latter which joined OBOR. Western Europe is more mixed but still I very much doubt you'll see a TikTok ban in Germany (and neither a Huawei one), and even in the UK I'm not sure if we'll see a software ban.

I think it's more likely we'll see a differentiated response. Strongest in the US and Australia maybe Japan (although that's an unclear case as well), and mixed to not at all in Europe.

On the global stage in general the situation is very different and for that reason alone China's global expansion has made sense. The few billion dollars they've invested in the US have given them extremely valuable knowledge.

"I personally have nothing to hide so I don't have a problem with intrusive surveillance" is not a nuclear or even hot take, in fact it's about the lukewarmest take I can think of.

For me it’s less about privacy concerns and more about not wanting to support an oppressive regime. Whether that be through financial boycotts or in the case of TikTok, through my data, which will help them perfect their products and give them a competitive edge which ultimately gives them more leverage to enact more oppressive and anti-democratic campaigns.

It's less of a matter of "I don't want the government to spy on me, the individual with a banal life" than a matter of "I don't want the government to spy on us, a society".

For instance, having a FB account enables FB to spy on non-FB users via shadow account - they will create shadow accounts based on faces they don't recognize in photos, and shadow friend profiles based on numbers stored in my phone.

A more insidious situation is that TikTok enables Bytedance (read: the CCP) to monitor political dissent and potentially shape cultural zeitgeist via content recommendation [1]

Many content recommendation algorithms use some sort of collaborative filtering algorithm, where if the system lacks data on one user's preferences they can infer it given a sufficient number of users with overlapping preferences and attributes [2]. So even though my data and life are not special, it does enable the government to find out more about people they are interested in.

[1] https://www.theguardian.com/technology/2019/sep/25/revealed-... [2] https://en.wikipedia.org/wiki/Collaborative_filtering

Seems like you’re misunderstanding the threat. You hinted at it with the collective for policing and advertising purposes. Handing over a picture of American life down to the individuals to the CCP gives them enormous power to create tailored disinformation campaigns that sow dissent, in addition to being a kind of real time probe into the mindsets of a foreign population. The culture wars that Trump tripled down on have partially been a product of foreign interference to create the seeds which the domestic audience happily latches onto and grows to division and in-fighting. Then there the issue that while you personally may have a boring life doesn’t mean that everyone does, and being able to harvest location at minimum from a variety of targets arbitrarily is a big coup for a foreign intelligence agency. These are just some of the possibilities.

> Your data and life are not special in any way. The government will never be interested in finding out about you, and if they really wanted to they could do it with ease, and definitely don't need an app to do it

And yet your data is valuable enough for multiple billion and trillion dollar companies to build massive infrastructure and markets to siphon your data away from you and sell it to the highest bidder.

We live in an unprecedented time where true mass surveillance is not only nearly free, it turns a profit. It costs a fraction of cent to record, transcribe and store the data you produce, so why not collect it? It's just a rounding error on some government agency's balance sheet.

The infrastructure is there. The data markets are there. The value of your data is there.

But most importantly, the evidence is there. In the last 20 years, governments have been caught using technology and private data to propagandize, suppress, persecute, blackmail, kill and even round people up into camps, like China does with its Uygher population[1].

Yes, you don't need an app to do this if you're a domestic government collecting your own citizens' data. But you might want an app if you're a government or business collecting data on foreign users.

Economic and military espionage and sabotage are incredibly valuable. If I was a nation-state, yes, I would like to have my app on foreign government workers' phones, the phones of foreign military members, the phones of professionals in competitive industries, the phones of family of friends of all of those people, and so on.

[1] https://www.reuters.com/investigates/special-report/muslims-...