To paraphrase The Hidden Complexity of Wishes - how would you ask an evil wish-granting genie to get your Mother out of a burning building? A gas pipe explosion will technically remove her body from the premises, but may not be quite what you had in mind. Such are the imperfections of language that all written law must ultimately be subject to intelligent human oversight. In their terms and conditions (T&C) software companies judiciously refer to themselves as ‘service’ providers. Are they?
Much of commercial law turns on the issue of accountability: when things go wrong, who pays? Broadly, three frameworks govern the question: contract law, common law for civil wrongs or ‘tort’ of negligence and statutory strict liability. Until the advent of mass production, contract law dominated. But industrialisation posed a new problem: privity of contract - the concept that no rights or obligations arising from one can apply to third parties - also meant no way for final buyers to make claims against producers since they rarely dealt with them directly. In Winterbottom v Wright (UK, 1842), Mr Winterbottom was denied compensation for injury when his poorly constructed mail coach collapsed because the Lords feared precedence: ‘the only safe rule’, they concluded, ‘is to confine the right to recover to those who enter into the contract’. It was not until courts introduced the concept of negligence, through pivotal cases such as Donoghue v Stevenson (UK, 1932) or MacPherson v Buick Motor Co (US, 1916), that buyers were allowed to get around privity. Yet negligence still requires plaintiffs to show ‘fault’, meaning blameworthiness and responsibility. In contrast, strict liability only requires that a tort occurred, regardless of culpability. If a product is defective, it does not matter producers followed the correct procedures: they are still liable. Product liability is the most important strict liability regime and it, crucially, applies only to manufacturers of products and not service providers. Doctors, lawyers, accountants, engineers - all still handle liability under the frameworks of contract law and negligence. It makes sense that software companies should prefer to trade under strict T&C (i.e. contracts) and avoid strict liability as much as possible.
Yet buyers may have more rights if software is considered a good, not a service. Replacement, reimbursement and damages are all easier to claim. In the US, courts have made the distinction on a case by case basis: is the software bespoke or mass-marketed? Did the buyer claim investment tax credits? Did the sales aspect predominate? Most notably, when bundled with hardware (as in laptops, Fitbits, Tesla cars), software will almost certainly be considered a ‘good’ and subject to Article 2 of the Uniform Commercial Code (UCC). Only this week, Microsoft paid $10 000 in a small claims court in California over a faulty Windows 10 upgrade. True, considering software a ‘good’ for the purpose of UCC has been more common than considering it a ‘product’ for the purpose of strict liability (‘good’ and ‘product’ are not perfect synonyms but the difference is still under debate). Yet invoking UCC makes invoking strict liability more likely. As the Internet of Things expands, expect more such rulings.
Already, in the UK, a key facet of the Consumer Rights Act 2015 (an EU directive implementation which updated and standardised previous consumer legislation) was to deal explicitly with ‘digital content’ and confer upon it the rights and obligations buyers would naturally expect. Under it, T&C are no longer binding if they ‘exclude or restrict the trader’s liability’. Ultimately, evil genies and T&C might be a necessary part of life, but mere crafty wording will not amount to genuine, binding consent. To agree, stop reading this now.
