THEY are creeping into your Twitter mentions, trying to make friends on Facebook, and liking your Instagram pics, all without you knowing. The legions of online bots 鈥� automated software designed to mimic humans 鈥� are growing, and politicians in the US and Europe have decided something needs to be done.
Last week, The Washington Post reported that as part of a bot crackdown, . While many bots are just a bit of fun, some are used for nefarious purposes. 鈥淭hey鈥檙e trying to trick people, whether into giving people more followers, allowing companies to inflate their valuations, or tricking people in politics,鈥� says Robert Hertzberg, a Californian state senator.
He is hoping to pass a new law 鈥� the Bolstering Online Transparency, or 鈥� that would require bots communicating with Californian residents to be identified on social networks. Bot creators would have to identify their products, and social networks would be responsible for removing any unidentified bots within 72 hours of their being highlighted by other users. Hertzberg hopes the Californian legislation will be a model for the wider US 鈥� and eventually, the rest of the world.
Advertisement
Automated Twitter activity has been recorded in the aftermath of school shootings, terrorist attacks and in the run-up to elections in at least 18 different countries, according to a , a non-governmental organisation focused on democracy. Such interference tries to shape public debate online. 鈥淲e can鈥檛 stop bots,鈥� says Hertzberg, 鈥渂ut we can at least put a label on them and make a process to bring them down.鈥�
His bill has already inspired one national politician to put bots in the firing line. Dianne Feinstein, who represents California in the US Senate, introduced a similar bill in late June.
鈥淭his bill is designed to help respond to Russia鈥檚 efforts to interfere in US elections through the use of social media bots, which spread divisive propaganda,鈥� she said in a statement.
The law would require social networks to disclose every bot operating on their platforms and is intended to stop 鈥渇ake social media bots鈥� being used by political campaigns.
鈥淰oters must be able to make informed judgements about the information they receive, and that includes knowing the origin and motives behind these pervasive and fake online personas,鈥� said Feinstein.
If the law is passed, social networks would have a year to flag up any bots on their platforms. Although what action would be taken against anyone flouting the rules has yet to be decided.
鈥淥nline disinformation is a new, multifaceted and fast-developing issue that requires immediate action鈥�
Meanwhile in Europe, politicians are also calling for social media firms to tackle the problem of automated accounts. As part of a wider movement against disinformation, the European Commission is developing a code of practice that it expects social networks to adhere to.
鈥淥nline disinformation is a new, multifaceted and fast-developing issue that requires immediate action,鈥� says Nathalie Vandystadt, a spokesperson for the commission. 鈥淪elf-regulation is considered the most appropriate way for online platforms to implement swift action to tackle this problem,鈥� she says, though the commission may introduce stronger regulation if sites don鈥檛 stick to the code.
The commission wants protections in place before the next European Parliament elections, due to take place in May 2019. 鈥淎ll efforts must be undertaken to counter disinformation tactics targeting election processes,鈥� says Vandystadt.
In fact, Europe may already have a law to deal with some bots. The General Data Protection Regulation (GDPR), which came into force in May, places restrictions on how people鈥檚 personal data can be used to make automated decisions. Because social media messages are personal data, a bot replying to your tweets should be covered by the law, says Jon Baines, chair of the UK鈥檚 National Association of Data Protection and Freedom of Information Officers.
鈥淏ots shouldn鈥檛 be taking, or be coded to take, decisions about us without our being aware,鈥� he says. 鈥淚f they are being used to make significant decisions about us, we should, in general, be allowed to refuse.鈥� That said, Baines admits the likelihood of Twitter bots being called to task for flouting GDPR rules is unlikely, although not impossible.
Friend or foe
But will identifying the bots make a difference to how we interact with them? A at the perceived differences between Twitter bots and human-controlled profiles found that users saw bots as slightly less competent and caring, although its findings were long before the rise of the bot army.
Without context as to who has created a bot and why, it can be difficult to discern what its goals are 鈥� and how you should treat the information it is providing. Laws that require bots to be identified should help with this, but may not be enough.
鈥淚 think it鈥檚 important to identify bots, but I think it鈥檚 more important to identify the people behind them,鈥� says Luciano Floridi, a digital ethicist at the Oxford Internet Institute. 鈥淭here is no such thing as independent bots unleashed on the web like some kind of alien life taking root. I think there should be accountability.鈥�
But not all bots are harmful, says Stefan Bohacek. He is creator of Botwiki, which catalogues friendly or helpful bots, such as those that create artwork or monitor changes to websites.
Bohacek says some bot-makers might not want to be identified for fear of reprisals when their bots tweet material someone doesn鈥檛 like. Or perhaps their accounts are benign art projects that rely on users believing they鈥檙e interacting with a human to avoid affecting their political viewpoints, so strict laws might go too far. 鈥淭he right balance would be to say this action has been automated, but I wouldn鈥檛 say we need to identify who鈥檚 automated it.鈥�
Hertzberg agrees it is important to strike a balance between cracking down on the bad bots while allowing space for positive experimentation. 鈥淲e don鈥檛 want to have a chilling effect on innovation and creativity,鈥� he says.
The Electronic Frontier Foundation, a digital rights group based in California, says Hertzberg鈥檚 bot legislation, as presently written, is flawed.
鈥淲e鈥檙e not opposed to bot labelling, but this bill is overbroad and poorly defined,鈥� says Jamie Lee Williams, staff attorney at the EFF.
The organisation in late May to outline its issues with his law. Like Bohacek, it raised concerns that bots used for research or artistic purposes would be affected, but it also highlighted wider problems.
The EFF says the BOT Act would require social networks to swiftly take down any unidentified bots flagged to them by a third party, then ask questions later. It worries that the law could be weaponised to take down benign bots that certain groups or individuals disagree with.
The group doesn鈥檛 go so far as to say bots themselves have the right to free speech, but it does say in its letter to Hertzberg that 鈥渄isclosure mandates would restrict and chill the speech鈥� of those behind non-harmful bots.
鈥淭here is no such thing as independent bots unleashed on the web like some kind of alien life鈥�
鈥淓lection interference and fraud are two areas in which bot activity is a huge problem and labelling would be helpful,鈥� says Williams, but it should be possible to address these problems without opening all bots up to attack. Removing strict takedown provisions from the law would be a helpful solution, she says.
Hertzberg doesn鈥檛 see the EFF鈥檚 qualms with his law. 鈥淚 heard their argument but can鈥檛 understand it,鈥� he says. 鈥淚f someone wants to have their voice, I鈥檓 not saying they can鈥檛. I just want to understand who the voice is from. All I鈥檓 suggesting is tell the truth about who you are.鈥�
The BOT Act is currently moving through the Californian legislature, and could become law later this year. But will it make a dent in the growing army? For that, we would need global agreement, which isn鈥檛 likely, says Floridi. 鈥淲e haven鈥檛 even achieved effective global regulation on the use of nuclear weapons,鈥� he says. 鈥淚鈥檇 be astonished if we were to reach it on this.鈥�
This article appeared in print under the headline 鈥淭he race to stop bots from taking over the world鈥�
