Respect can be a rare commodity online, depending on what Twitter threads or Facebook groups you are in. But the word was used firmly by the UK government last month when it said it would resist attempts to water down its proposed powers over the internet.
The mannered language belies a wolf in sheep’s clothing, say critics of the online safety bill. This much-contested legislation returns to on 12 July, and MPs parliament made it clear this week that they think the culture secretary will have too much power over the internet as it stands.
Julian Knight, the Conservative MP who chairs the digital, culture, media and sport committee, will have warned that the secretary of state influence over Ofcom, the independent regulator charged with implementing the act. He called for the removal of clauses that would allow Nadine Dorries, still the culture secretary at the time of publication, to order Ofcom to change codes of practice, including on dealing with terrorist and child sexual exploitation content, before parliament considers them.
“A free media depends on ensuring the regulator is free from the threat of day-to-day interference from the executive,” said Knight. “The government will still have an important role in setting the direction of travel, but Ofcom must not be constantly peering over its shoulder answering to the whims of a backseat-driving secretary of state.”
The government was polite with its hard no. Speaking to a committee of MPs scrutinizing the bill last month, the digital minister, Chris Philp, said the government would “respectfully resist” attempts to water down the secretary of state’s powers.
The government won’t move on that point, but it is introducing changes nonetheless.
The bill places a duty of care on tech firms – or rather, platforms that produce user-generated content such as social media giants, as well as big search engines including Google – to protect users from harmful content. That duty of care is broadly split into three parts: limiting the spread of illegal content such as child sexual abuse images and terrorist material; ensuring children are not exposed to harmful or inappropriate content; and, for the big platforms such as Facebook, Twitter and TikTok, protecting adults from legal but harmful content (such as cyberbullying and eating disorder-related material).
The legislation will be overseen by Ofcom, which will be able to impose fines of £18m or 10% of a company’s global turnover for breaches of the act. In extreme cases, it can also block websites or apps. On Wednesday Ofcom published its roadmap for implementing the act, including a focus on tackling illegal content within the first 100 days of the legislation being implemented.
Here is a quick precis of what changes to expect as the bill enters its next stage. It should become law by the end of the year or in early 2023, depending on how it does in the House of Lords, which is bound to have a few issues with it.
Ch-ch-changes: confirmed amendments
The government is introducing some amendments in time for the report stage on 12 July, with another batch to be announced shortly after. Under one confirmed change, tech firms will be required to shield internet users from state-sponsored disinformation that poses a threat to UK society and democracy. This is a tightening of existing proposals on disinformation in the bill, which already require tech firms to take action on state-sponsored disinformation that harms individuals – such as threats to kill.
Another confirmed amendment is equally incremental. A clause in the bill aimed at end-to-end encrypted services already gives Ofcom the power to require those platforms to adopt “accredited technology” to detect child sexual abuse and exploitation [CSEA] content. If that doesn’t work, then they must use their “best endeavors” to develop or deploy new technology to spot and remove CSEA. This move appears to be aimed at Mark Zuckerberg’s plans to introduce end-to-end encryption on Facebook Messenger and Instagram.
First do no harm: what’s expected
At the committee stage, Philp confirmed that, one way or another, the government will bring in an offence on the deliberate sending of flashing images to incite epileptic fits. However, it might not be in the safety bill.
He also said that “in due course” the government will publish a list of “priority harms” to adults, indicating a change to the original plan of publishing them becomes after the bill law. These are the harms – nasty but not criminal – that fall below the threshold of illegality that must be addressed by platforms, which are expected to include self-harm, harassment and eating disorders. There is concern that this will turn the bill into a censors’ charter where tech firms turn against content that exists in a gray area of acceptability, like satire.
William Perrin, a trustee of the Carnegie UK Trust charity, wants the government to go further and publish those priority harms in the amended bill so that MPs can debate them before they become law. “Regulation of the media should be independent of the executive,” he says. “The government needs to give up the power to define harmful but not illegal content and instead hammer it out in parliament.”
Widening the criminal landscape: other changes
The “priority harms” clause applies to so-called category 1 tech firms, the big hitters like Facebook, Instagram, Twitter, YouTube and TikTok. There are calls to expand that list to edgier platforms such as 4chan and BitChute, which definitely contain harmful content.
Philp also told MPs last month that he would consider calls to add more criminal offenses to the list of illegal content – related to real-world criminality – that must be tackled by all firms within the scope of the bill. Trafficking and modern slavery were among the criminal offenses that MPs want included. At the moment, the “priority offenses” written into the bill include selling firearms illegally and threats to kill.
If you want to read the complete version of the newsletter, please subscribe to receive TechScape in your inbox every Wednesday.