When Is a Twitter Bot Not a Twitter Bot?

The United States of America. Russia. If you have been following the news, you know the relationship between these two countries can be complicated. But one thing is not complicated enough: the algorithms used by Twitter to identify fake bots that were created by Russians to troll Americans and destabilize their faith in the US democratic process. We know this because Wired has revealed that several real-life Americans were included in a list of suspected Russian Twitter bots, as identified by Twitter’s algorithms and then handed to the press by politicians.

It is hardly surprising that the staff working for Twitter will make mistakes. Theirs is a business which seems defined by the capacity to wield enormous social power and generate huge amounts of money without owning anything of worth or accepting any legal responsibility for anything. You cannot expect the best insights from a business that took 12 years to deliver their first quarterly profit but still has to remove millions of fake accounts. Even one of their board members was caught inflating her popularity by buying fake followers. We can only wonder what the true profitability of Twitter would be if all the bogus accounts were removed from the system. Nevertheless, they may have incorrectly included as many as 20 real users in a list of less than 4,000 suspected bots. Mistakenly denying access to 1 person in every 200 may not sound like much, but imagine if a telco wrongly disconnected that many phone users, or an ISP denied internet access to that many genuine customers. And in this instance Twitter was trying really hard to identify the bots – how well do they perform when the stakes are less crucial than providing the government with information meant to defend national security?

There are many ethical issues when large communications corporations misidentify human customers and treat them like machines. The reputations of some real people have been damaged because grandstanding politicians like Congressman Adam Schiff decided to publish the suspected Twitter accounts as part of their campaign to make a broader political point. Beyond the reputation harm, taking away the accounts of these real users has denied them a potentially powerful channel for free speech. The usual retort is that Twitter is a private enterprise, so there is no public right to free speech on that platform. However, that strict division between public rights and private services has been fatally undermined by the legal decision that every Twitter user has the right to troll the US President, which presumably means the President cannot even block an account which behaves just like a Russian bot. Only a topsy-turvy logic would conclude everybody has the legal right to troll public officials on Twitter although nobody has any legal right to demand access to Twitter in the first place. To be fair, Jameel Jaffer, executive director of the Knight First Amendment Institute, the free speech advocacy group that won the case which stops Trump from blocking Twitter trolls, also questioned whether Schiff and the other politicians who published Twitter’s list of suspected bots…

…took the steps [they] ought to have taken before publishing information that [they] should have known would be deeply damaging to the people who were named.

Lives can be shattered by the denial or abuse of a communications service. Our world is more connected than ever. The greater the number of people who rely on massively connected communications services, the greater the loss suffered by any individual who is denied access. Furthermore, we are training ourselves for a world where the data being gathered will be increasingly used for ever more serious decisions to be made purely by machines, without human oversight. We must put aside the hype and realize current technology is not good enough to reliably distinguish between real people and machines.

The purpose of a Turing test is to encourage the building of machines that will be indistinguishable from real people, but our society has already started to rely on faulty technology which mistakes real people for machines. We must stop the rush towards exploitation of immature tech. One mistake in every 200 may not sound like many to a President or a CEO – they can win elections and generate profits from the other 199 people – but each of those mistakes is endured by an actual person. We can all imagine the dire consequences of being that one person who is wrongly excluded, or has been subjected to the wrong decisions.

Some of our business and political leaders lack the intellect, knowledge or seriousness to stop the damage before it is too late. Those of us working in the communications industry, and especially those of us conscious of the risks taken by the communications industry, must do all we can to protect ordinary people, even if nobody else will.

You can read the full Wired story about Twitter’s flawed list of Twitter bots by looking here.

Eric Priezkalns
Eric Priezkalns
Eric is the Editor of Commsrisk. Look here for more about the history of Commsrisk and the role played by Eric.

Eric is also the Chief Executive of the Risk & Assurance Group (RAG), a global association of professionals working in risk management and business assurance for communications providers.

Previously Eric was Director of Risk Management for Qatar Telecom and he has worked with Cable & Wireless, T‑Mobile, Sky, Worldcom and other telcos. He was lead author of Revenue Assurance: Expert Opinions for Communications Providers, published by CRC Press. He is a qualified chartered accountant, with degrees in information systems, and in mathematics and philosophy.