I sometimes wonder to what extent the comparative improvement in the abilities of Artificial Intelligence, relative to human intelligence, is as a result of humans becoming more stupid.
\ This idea came to mind recently when someone shared an article with me from those in our own field, software engineering. With social media regulation upcoming in the UK, the admin of the software engineering community site lobste.rs believed the most rational approach to take would be to ask their own members for legal advice.
\ Whilst I would perhaps go and speak to someone who is qualified in one of the three legal jurisdictions in the UK, who am I to judge this non-conventional approach to gaining legal advice?
\ Naturally, this should be something of interest to me, as I am currently doing significant work at the intersection of socio-techno-legal matters.
\ I have been asked to give my views in private gatherings when legislators were discussing the Online Safety Act and have helped draft other legislation, I have been instructed as an expert witness on technology matters in court cases and, whilst I am not yet a lawyer, I am on an accelerated graduate-entry program to be one in Scotland (whose distinctive legal system is broadly a mix of the two great legal traditions in the world; English common law and Roman law).
\ I generally spend an equal amount of time talking to lawyers nowadays as software engineers. In fact, I drafted my first particulars of claim in a real-life lawsuit just recently (I’ll speak more about how that went below).
\ So, of course, I was naturally intrigued to be sent an insightful online discussion on the Online Safety Act by like-minded individuals with an interest in addressing such issues in the public interest. Reading the comments, I was simply horrified by the amount of confidently presented misinformation that was unchallenged. This was seemingly driven by a desire to gain validation by criticizing the work of well-intentioned legislation draftspersons without any understanding of the domain they operate within.
\ Whilst these conversations are hard to have, I do not mean this article to come across as a ‘diss’, but I write this to highlight how far too often intellectual discussions in tech go from being about the arguments in one’s brains to a contest of measuring what’s in their underwear.
Debunking the Diatribe Interpreting Law as SoftwareThe second comment from the top of the Lobsters post represents much of the discussion. It comes from someone whose expertise of the various legal systems in place in the UK seems to be “having worked for a MEP [Member of the European Parliament]”, seemingly in the Czech Republic). They write:
“Well, I’ve started to read the actual law and OH MY GOD, what a mess! Are all UK laws written in this impenetrable style? … Oh god, kill me now. Yes, captain obvious, that’s what legislation is supposed to do. I am holding my breath, I am literally reading the law here, the suspense is killing me, do tell, what are the duties of various kinds of search service providers?”
\ This disregards the fact that law is not software - nowadays the courts look to the purpose of legislation rather than purely textual interpretation. As was restated in [70] of Uber BV v Aslam [2021] UKSC 5:
“The modern approach to statutory interpretation is to have regard to the purpose of a particular provision and to interpret its language, so far as possible, in the way which best gives effect to that purpose…”
Reading the DefinitionsOne comment puts forward the following unchallenged statement:
‘The act requires that websites have policies in place to protect people from “harmful” or illegal content while being entirely vague as to what that means, what counts as “harmful” appears to have been kept intentionally vague so as to become anything the prosecutor wants it to be. It genuinely feels as though this is the Thought Police Act.’
\ Yet, this could be found to be simply untrue by reading the Act. At S60-62, “content that is harmful to children”, “primary priority content that is harmful to children” and “priority content that is harmful to children” are clearly defined and at S59 yet more clear definitions can be found for “illegal content”, “priority offence”, “terrorism content”, “CSEA content” and “priority illegal content.”
\ This is signposted numerous times throughout the Act, and such definitions are under their own Chapter in the relevant Part of the Act.
This Takes the Cake…Another unchallenged statement reads:
“It is in general that laws keep this vague and courts decide the rest. This is because morale and knowledge works faster than lawmaking. So, yes, it is kept at liberty to the prosecutor, with the prosecutor to work out the details and set and evolve policy. That is something that the Ofcom is currently doing, but in a pretty poor way.”
\ To briefly address some of the numerous factual errors in turn here:
“It is in general that laws keep this vague and courts decide the rest.” - This is nonsense. The Codes of Practice which OFCOM draft must be submitted to the Secretary of State before being approved by Parliament for them to have legal effect. S178 of this legislation places a duty on the Secretary of State to conduct a review of the regulatory framework provided for in this Act within two to five years of the Act coming into force.
\
“This is because morale and knowledge works faster than lawmaking.” - Lawmaking can work very fast. We have secondary legislation like Regulations (and even tertiary legislation sometimes), which can change the law without requiring whole new Acts of Parliament - this was used for postcode-specific COVID-19 lockdowns in England and to restrict airspace in the UK as needed. They can be “made affirmative” where the primary legislation allows it, meaning it’s the law right after a government minister has signed it, before Parliament has even considered it.
\
“So, yes, it is kept at liberty to the prosecutor, with the prosecutor to work out the details and set and evolve policy.” - Prosecution is something that happens in criminal law and in Scotland, only the Crown Office and Procurator Fiscal Service (COPFS) can prosecute people*, not OFCOM. In England, whilst technically anyone can bring a private prosecution, they must follow the Code for Crown Prosecutors, meaning the lawyers prosecuting "are independent from persons or agencies that are not part of the prosecution decision-making process" (and following the Post Office Horizon IT scandal this may well be restricted further). Furthermore, it is of course ultimately up to a court to pronounce the law, not a prosecutor.
\ More fundamentally, the maxim nullum crimen sine lege (no crime without law) is enshrined in Article 7 of the European Convention on Human Rights (alongside multiple other civil liberty safeguards), which, per the Human Rights Act, courts must give regard to. Even before the enactment of the Human Rights Act, when the defense to the offense of marriage was removed from the offense of rape in English & Welsh law, following the position that the Scottish Courts had reached, Lord Lane justified his decision by stating “This is not the creation of a new offence, it is the removal of a common law fiction…” (R v R [1991] UKHL 12).
\ * Footnote: Extremely rarely and only in special circumstances, private criminal prosecution can happen in Scotland with permission of the High Court of Justiciary, which has only happened twice in the last century.
About Arrogance, Not LawI am not commenting here on the merits of the Online Safety Act, but I want to focus on the fact that so much factually untrue material is being shared, even though critical thinking and background reading would have made the nonsense clear to the authors. However, what I’ve shared here is just a fraction of the nonsense that was shared with unquestioning confidence in these comments, without diving into the misconceptions of international law or the consequences of breaching the legislation.
\ Even the less diabolical comments contain information that can simply be dispelled by reading Ofcom’s short: “Quick guide to illegal content codes of practice”. Moreover, this misinformation remains unchallenged despite being easily identifiable as untrue.
\ It is worth stressing this isn’t a one-off; it isn’t uncommon to come across such misinformation in a variety of subjects, roll my eyes, and move on (or be sent it by others to which we have a good chuckle). Yet, the impact can be worse for those who are unable to evaluate the merit of such information when presented with hubris.
When Software Meets LawFor the case I mentioned earlier to be filed, I found myself at a courthouse having to apologetically run the Clerk through a rarely used provision in a Practice Direction to allow a pursuer to submit a claim by hand. The reason why is the online submission system didn’t seem to be programmed with the awareness Scotland is a separate legal jurisdiction to England & Wales, meaning the claim would be routed to an office in England when the arguments in this claim were rooted in the Scots Law of contract.
\ Post wasn’t a risk I was willing to take, there have been cases where a claim was received 88 seconds after the deadline and the ruling was there was “no margin for error” (Beasley v National Grid [2008] EWCA Civ 742). I didn’t want to leave this to a last-minute dash at the mercy of the postal service.
\ I ended up solving the problem by finding a rarely used procedure in a Practice Direction that allows a claim to be presented by hand at one of four courthouses in Scotland (though this method is non-obvious to non-lawyers because it is not listed on the GOV.UK website).
\ After my apologetic explanation, the Clerk returned with a copy of the front page stamped “RECEIVED 06 JAN 2025” (photographed).
\ I wonder if this situation would have been avoided if the developers who built the software either took the time to understand the fundamentals of what they were building, spoke to Scottish lawyers, or gathered the requirements right.
\ Whenever I’ve dealt with court and tribunal staff, they have always been profoundly empathetic and deeply understanding, so I have no reason to assume this isn’t true of the technology teams building this submission flow - so the engineering issues very much seem as if they come from a lack of understanding and are unintentional. However, that doesn’t mean we shouldn’t try to engineer systems that better match the world they seek to model.
Improving Emotional IntelligenceThe book, When Prophecy Fails, says of cognitive dissonance:
"Tell him you disagree and he turns away. Show him facts or figures and he questions your sources. Appeal to logic and he fails to see your point."
\ An example of this could be seen recently when Martin Fowler and others shared fundamentally untrue information about statistical techniques.
\ However, as the case of the Online Safety Act demonstrates, this goes beyond merely defending tightly held opinions. Whilst some may need to espouse such ignorance from a desire for validation or as an ego-defense mechanism, it does no service to our profession or other professionals who we increasingly need to ask to share their world with ours.
\ It is important to recognize those who act in such a way are in a minority in our profession, but they do not help us.
\ A few months ago, when I told a law professor that I came from a software engineering background, his response was to say: “Oh, you poor thing.” At the time, I laughed it off. However, seeing the perspective of software from the law, I can come closer to understanding what he was getting at.
\ Whilst it’s true that to some extent our Emotional Intelligence (EI) can be dictated by nature, a 2019 Randomised Control Trial study found consistent with previous studies “that EI can be improved within business environments”. However, for these skills to be developed, incentives must be in place in our industry to develop them.
\ Maybe it’s time for us to place greater value on self-awareness and emotional awareness in software engineering.
All Rights Reserved. Copyright , Central Coast Communications, Inc.